Does Linux support powering down SATA drives?

Does Linux support powering down SATA drives?

[Marc Perkel]

Trying to save power consumption. I have a backup drive that is used 
only once a day to back up the main drive. So - why should I run it more 
that 10 minutes a day? What I'd like to do is keep it in an off state 
and then at night power it on, mount it up, do the backup, unmount it, 
and shut it down. Can I do that?

Re: Does Linux support powering down SATA drives?

[Alistair John Strachan]

On Saturday 19 November 2005 19:01, Alan Cox wrote:
> On Sad, 2005-11-19 at 09:41 -0800, Marc Perkel wrote:
> > Trying to save power consumption. I have a backup drive that is used
> > only once a day to back up the main drive. So - why should I run it more
> > that 10 minutes a day? What I'd like to do is keep it in an off state
> > and then at night power it on, mount it up, do the backup, unmount it,
> > and shut it down. Can I do that?
>
> SATA not yet, USB you could however.

Or PATA, of course. I switch off two of my HDs 4 minutes after last use with 
the commands:

hdparm -S 48 /dev/hde
hdparm -S 48 /dev/hdg

Isn't there a passthru patch in the works to let commands, such as the one 
required for suspend, through to a SATA device?

-- 
Cheers,
Alistair.

'No sense being pessimistic, it probably wouldn't work anyway.'
Third year Computer Science undergraduate.
1F2 55 South Clerk Street, Edinburgh, UK.

Re: Does Linux support powering down SATA drives?

[Alan Cox]

On Sad, 2005-11-19 at 09:41 -0800, Marc Perkel wrote:
> Trying to save power consumption. I have a backup drive that is used 
> only once a day to back up the main drive. So - why should I run it more 
> that 10 minutes a day? What I'd like to do is keep it in an off state 
> and then at night power it on, mount it up, do the backup, unmount it, 
> and shut it down. Can I do that?


SATA not yet, USB you could however.

Re: Does Linux support powering down SATA drives?

[Marc Perkel]

Alistair John Strachan wrote:

>On Saturday 19 November 2005 19:01, Alan Cox wrote:
>  
>
>>On Sad, 2005-11-19 at 09:41 -0800, Marc Perkel wrote:
>>    
>>
>>>Trying to save power consumption. I have a backup drive that is used
>>>only once a day to back up the main drive. So - why should I run it more
>>>that 10 minutes a day? What I'd like to do is keep it in an off state
>>>and then at night power it on, mount it up, do the backup, unmount it,
>>>and shut it down. Can I do that?
>>>      
>>>
>>SATA not yet, USB you could however.
>>    
>>
>
>Or PATA, of course. I switch off two of my HDs 4 minutes after last use with 
>the commands:
>
>hdparm -S 48 /dev/hde
>hdparm -S 48 /dev/hdg
>
>Isn't there a passthru patch in the works to let commands, such as the one 
>required for suspend, through to a SATA device?
>
>  
>

So - why isn't there more SATA support. Seems like this and SMART aren't 
supported. What's up with that? Why is SATA harder than IDE?

-- 
Marc Perkel - marc@perkel.com

Spam Filter: http://www.junkemailfilter.com
    My Blog: http://marc.perkel.com

USB storage -> Oops (2.6.14.2)

[JaniD++]

Hello list,

I have pugged in two USB storage.

1. WD 200GB
Nov 19 19:23:57 192.168.2.50 SCSI device sda: 268435455 512-byte hdwr
sectors (137439 MB)
(128GB???)

2. Maxtor 40GB
Nov 19 19:24:02 192.168.2.50 SCSI device sdb: 80293248 512-byte hdwr sectors
(41110 MB)


The full log is here:
(Syslog+netconsole)

Nov 19 19:23:51 192.168.2.50 kernel: usb 1-3: new high speed USB device
using ehci_hcd and address 2
Nov 19 19:23:51 192.168.2.50 kernel: scsi0 : SCSI emulation for USB Mass
Storage devices
Nov 19 19:23:56 192.168.2.50   Vendor:
Nov 19 19:23:56 192.168.2.50 U
Nov 19 19:23:56 192.168.2.50 S
Nov 19 19:23:56 192.168.2.50 B
Nov 19 19:23:56 192.168.2.50
Nov 19 19:23:56 192.168.2.50 2
Nov 19 19:23:56 192.168.2.50 .
Nov 19 19:23:56 192.168.2.50 0
Nov 19 19:23:56 192.168.2.50
Nov 19 19:23:56 192.168.2.50   Model:
Nov 19 19:23:56 192.168.2.50 S
Nov 19 19:23:56 192.168.2.50 t
Nov 19 19:23:56 192.168.2.50 o
Nov 19 19:23:56 192.168.2.50 r
Nov 19 19:23:56 192.168.2.50 a
Nov 19 19:23:56 192.168.2.50 g
Nov 19 19:23:56 192.168.2.50 e
Nov 19 19:23:56 192.168.2.50
Nov 19 19:23:56 192.168.2.50 D
Nov 19 19:23:56 192.168.2.50 e
Nov 19 19:23:56 192.168.2.50 v
Nov 19 19:23:56 192.168.2.50 i
Nov 19 19:23:56 192.168.2.50 c
Nov 19 19:23:56 192.168.2.50 e
Nov 19 19:23:56 192.168.2.50
Nov 19 19:23:56 192.168.2.50
Nov 19 19:23:56 192.168.2.50   Rev:
Nov 19 19:23:56 192.168.2.50 0
Nov 19 19:23:56 192.168.2.50 1
Nov 19 19:23:56 192.168.2.50 0
Nov 19 19:23:56 192.168.2.50 0
Nov 19 19:23:56 192.168.2.50
Nov 19 19:23:56 192.168.2.50   Type:   Direct-Access
Nov 19 19:23:57 192.168.2.50                  ANSI SCSI revision: 00
Nov 19 19:23:57 192.168.2.50
Nov 19 19:23:57 192.168.2.50 kernel:   Vendor: USB 2.0   Model: Storage
Device    Rev: 0100
Nov 19 19:23:57 192.168.2.50 kernel:   Type:   Direct-Access
ANSI SCSI revision: 00
Nov 19 19:23:57 192.168.2.50 SCSI device sda: 268435455 512-byte hdwr
sectors (137439 MB)
Nov 19 19:23:57 192.168.2.50 sda: assuming drive cache: write through
Nov 19 19:23:57 192.168.2.50 kernel: SCSI device sda: 268435455 512-byte
hdwr sectors (137439 MB)
Nov 19 19:23:57 192.168.2.50 kernel: sda: assuming drive cache: write
through
Nov 19 19:23:57 192.168.2.50 SCSI device sda: 268435455 512-byte hdwr
sectors (137439 MB)
Nov 19 19:23:57 192.168.2.50 sda: assuming drive cache: write through
Nov 19 19:23:57 192.168.2.50 kernel: SCSI device sda: 268435455 512-byte
hdwr sectors (137439 MB)
Nov 19 19:23:57 192.168.2.50 kernel: sda: assuming drive cache: write
through
Nov 19 19:23:57 192.168.2.50 kernel:  sda: sda1 sda2
Nov 19 19:23:57 192.168.2.50 Attached scsi disk sda at scsi0, channel 0, id
0, lun 0
Nov 19 19:23:57 192.168.2.50 kernel: Attached scsi disk sda at scsi0,
channel 0, id 0, lun 0
Nov 19 19:23:57 192.168.2.50 Attached scsi generic sg0 at scsi0, channel 0,
id 0, lun 0,  type 0
Nov 19 19:23:57 192.168.2.50 kernel: Attached scsi generic sg0 at scsi0,
channel 0, id 0, lun 0,  type 0
Nov 19 19:23:57 192.168.2.50 kernel: usb 1-1: new high speed USB device
using ehci_hcd and address 3
Nov 19 19:23:57 192.168.2.50 kernel: scsi1 : SCSI emulation for USB Mass
Storage devices
Nov 19 19:24:02 192.168.2.50   Vendor:
Nov 19 19:24:02 192.168.2.50 U
Nov 19 19:24:02 192.168.2.50 S
Nov 19 19:24:02 192.168.2.50 B
Nov 19 19:24:02 192.168.2.50
Nov 19 19:24:02 192.168.2.50 2
Nov 19 19:24:02 192.168.2.50 .
Nov 19 19:24:02 192.168.2.50 0
Nov 19 19:24:02 192.168.2.50
Nov 19 19:24:02 192.168.2.50   Model:
Nov 19 19:24:02 192.168.2.50 S
Nov 19 19:24:02 192.168.2.50 t
Nov 19 19:24:02 192.168.2.50 o
Nov 19 19:24:02 192.168.2.50 r
Nov 19 19:24:02 192.168.2.50 a
Nov 19 19:24:02 192.168.2.50 g
Nov 19 19:24:02 192.168.2.50 e
Nov 19 19:24:02 192.168.2.50
Nov 19 19:24:02 192.168.2.50 D
Nov 19 19:24:02 192.168.2.50 e
Nov 19 19:24:02 192.168.2.50 v
Nov 19 19:24:02 192.168.2.50 i
Nov 19 19:24:02 192.168.2.50 c
Nov 19 19:24:02 192.168.2.50 e
Nov 19 19:24:02 192.168.2.50
Nov 19 19:24:02 192.168.2.50
Nov 19 19:24:02 192.168.2.50   Rev:
Nov 19 19:24:02 192.168.2.50 0
Nov 19 19:24:02 192.168.2.50 1
Nov 19 19:24:02 192.168.2.50 last message repeated 2 times
Nov 19 19:24:02 192.168.2.50
Nov 19 19:24:02 192.168.2.50   Type:   Direct-Access
Nov 19 19:24:02 192.168.2.50                  ANSI SCSI revision: 00
Nov 19 19:24:02 192.168.2.50
Nov 19 19:24:02 192.168.2.50 SCSI device sdb: 80293248 512-byte hdwr sectors
(41110 MB)
Nov 19 19:24:02 192.168.2.50 sdb: assuming drive cache: write through
Nov 19 19:24:02 192.168.2.50 SCSI device sdb: 80293247 512-byte hdwr sectors
(41110 MB)
Nov 19 19:24:02 192.168.2.50 sdb: assuming drive cache: write through
Nov 19 19:24:02 192.168.2.50 kernel:   Vendor: USB 2.0   Model: Storage
Device    Rev: 0111
Nov 19 19:24:02 192.168.2.50 kernel:   Type:   Direct-Access
ANSI SCSI revision: 00
Nov 19 19:24:02 192.168.2.50 kernel: SCSI device sdb: 80293248 512-byte hdwr
sectors (41110 MB)
Nov 19 19:24:02 192.168.2.50 kernel: sdb: assuming drive cache: write
through
Nov 19 19:24:02 192.168.2.50 kernel: SCSI device sdb: 80293247 512-byte hdwr
sectors (41110 MB)
Nov 19 19:24:02 192.168.2.50 kernel: sdb: assuming drive cache: write
through
Nov 19 19:24:02 192.168.2.50 kernel:  sdb: sdb1
Nov 19 19:24:02 192.168.2.50 Attached scsi disk sdb at scsi1, channel 0, id
0, lun 0
Nov 19 19:24:02 192.168.2.50 kernel: Attached scsi disk sdb at scsi1,
channel 0, id 0, lun 0
Nov 19 19:24:02 192.168.2.50 Attached scsi generic sg1 at scsi1, channel 0,
id 0, lun 0,  type 0
Nov 19 19:24:02 192.168.2.50 kernel: Attached scsi generic sg1 at scsi1,
channel 0, id 0, lun 0,  type 0
Nov 19 19:24:02 192.168.2.50 Unable to handle kernel NULL pointer
dereference
Nov 19 19:24:02 192.168.2.50  at virtual address 00000000
Nov 19 19:24:02 192.168.2.50  printing eip:
Nov 19 19:24:02 192.168.2.50 c03df86c
Nov 19 19:24:02 192.168.2.50 *pde = 006d1001
Nov 19 19:24:02 192.168.2.50 Oops: 0000 [#1]
Nov 19 19:24:02 192.168.2.50 SMP
Nov 19 19:24:02 192.168.2.50
Nov 19 19:24:02 192.168.2.50 Modules linked in:
Nov 19 19:24:02 192.168.2.50  netconsole
Nov 19 19:24:02 192.168.2.50
Nov 19 19:24:02 192.168.2.50 CPU:    3
Nov 19 19:24:02 192.168.2.50 EIP:    0060:[<c03df86c>]    Not tainted VLI
Nov 19 19:24:02 192.168.2.50 EFLAGS: 00010286   (2.6.14.2)
Nov 19 19:24:02 192.168.2.50 EIP is at scsi_run_queue+0x12/0xbc
Nov 19 19:24:02 192.168.2.50 eax: 00000000   ebx: db480ca4   ecx: 00000001
edx: db480ca4
Nov 19 19:24:02 192.168.2.50 esi: f49bdb00   edi: 00000246   ebp: c35bfe88
esp: c35bfe74
Nov 19 19:24:02 192.168.2.50 ds: 007b   es: 007b   ss: 0068
Nov 19 19:24:02 192.168.2.50 Process ksoftirqd/3 (pid: 12,
threadinfo=c35be000 task=c3550030)
Nov 19 19:24:02 192.168.2.50
Nov 19 19:24:02 192.168.2.50 Stack:
Nov 19 19:24:02 192.168.2.50 00000060
Nov 19 19:24:02 192.168.2.50 00000282
Nov 19 19:24:02 192.168.2.50 db480ca4
Nov 19 19:24:02 192.168.2.50 f49bdb00
Nov 19 19:24:02 192.168.2.50 00000246
Nov 19 19:24:02 192.168.2.50 c35bfe98
Nov 19 19:24:02 192.168.2.50 c03df989
Nov 19 19:24:03 192.168.2.50 db480ca4
Nov 19 19:24:03 192.168.2.50
Nov 19 19:24:03 192.168.2.50
Nov 19 19:24:03 192.168.2.50 e6ba79dc
Nov 19 19:24:03 192.168.2.50 c35bfebc
Nov 19 19:24:03 192.168.2.50 c03dfa7e
Nov 19 19:24:03 192.168.2.50 f49bdb00
Nov 19 19:24:03 192.168.2.50 00000000
Nov 19 19:24:03 192.168.2.50 00000024
Nov 19 19:24:03 192.168.2.50 db480ca4
Nov 19 19:24:03 192.168.2.50 f49bdb00
Nov 19 19:24:03 192.168.2.50
Nov 19 19:24:03 192.168.2.50
Nov 19 19:24:03 192.168.2.50 00040000
Nov 19 19:24:03 192.168.2.50 e6ba79dc
Nov 19 19:24:03 192.168.2.50 c35bff08
Nov 19 19:24:03 192.168.2.50 c03dff01
Nov 19 19:24:03 192.168.2.50 f49bdb00
Nov 19 19:24:03 192.168.2.50 00000000
Nov 19 19:24:03 192.168.2.50 00000024
Nov 19 19:24:03 192.168.2.50 00000001
Nov 19 19:24:03 192.168.2.50
Nov 19 19:24:03 192.168.2.50 Call Trace:
Nov 19 19:24:03 192.168.2.50  [<c0103bf2>]
Nov 19 19:24:03 192.168.2.50 show_stack+0x9a/0xd0
Nov 19 19:24:03 192.168.2.50
Nov 19 19:24:03 192.168.2.50  [<c0103db2>]
Nov 19 19:24:03 192.168.2.50 show_registers+0x16a/0x1fa
Nov 19 19:24:03 192.168.2.50
Nov 19 19:24:03 192.168.2.50  [<c0103fc3>]
Nov 19 19:24:03 192.168.2.50 die+0xfa/0x17c
Nov 19 19:24:03 192.168.2.50
Nov 19 19:24:03 192.168.2.50  [<c05557bc>]
Nov 19 19:24:03 192.168.2.50 do_page_fault+0x37c/0x6da
Nov 19 19:24:03 192.168.2.50
Nov 19 19:24:03 192.168.2.50  [<c01038ab>]
Nov 19 19:24:03 192.168.2.50 error_code+0x4f/0x54
Nov 19 19:24:03 192.168.2.50
Nov 19 19:24:03 192.168.2.50  [<c03df989>]
Nov 19 19:24:03 192.168.2.50 scsi_next_command+0x20/0x26
Nov 19 19:24:03 192.168.2.50
Nov 19 19:24:03 192.168.2.50  [<c03dfa7e>]
Nov 19 19:24:03 192.168.2.50 scsi_end_request+0xb6/0xf9
Nov 19 19:24:03 192.168.2.50
Nov 19 19:24:03 192.168.2.50  [<c03dff01>]
Nov 19 19:24:03 192.168.2.50 scsi_io_completion+0x2c0/0x4ef
Nov 19 19:24:03 192.168.2.50
Nov 19 19:24:03 192.168.2.50  [<c03e0398>]
Nov 19 19:24:03 192.168.2.50 scsi_generic_done+0x3b/0x47
Nov 19 19:24:03 192.168.2.50
Nov 19 19:24:03 192.168.2.50  [<c03db491>]
Nov 19 19:24:03 192.168.2.50 scsi_finish_command+0x83/0x97
Nov 19 19:24:03 192.168.2.50
Nov 19 19:24:03 192.168.2.50  [<c03db359>]
Nov 19 19:24:03 192.168.2.50 scsi_softirq+0xb8/0x134
Nov 19 19:24:03 192.168.2.50
Nov 19 19:24:03 192.168.2.50  [<c0123ec2>]
Nov 19 19:24:03 192.168.2.50 __do_softirq+0x72/0xdc
Nov 19 19:24:03 192.168.2.50
Nov 19 19:24:03 192.168.2.50  [<c0123f63>]
Nov 19 19:24:03 192.168.2.50 do_softirq+0x37/0x39
Nov 19 19:24:03 192.168.2.50
Nov 19 19:24:03 192.168.2.50  [<c012447c>]
Nov 19 19:24:03 192.168.2.50 ksoftirqd+0xa5/0xfa
Nov 19 19:24:03 192.168.2.50
Nov 19 19:24:03 192.168.2.50  [<c01330d7>]
Nov 19 19:24:03 192.168.2.50 kthread+0xb1/0xb5
Nov 19 19:24:03 192.168.2.50
Nov 19 19:24:03 192.168.2.50  [<c0101145>]
Nov 19 19:24:03 192.168.2.50 kernel_thread_helper+0x5/0xb
Nov 19 19:24:03 192.168.2.50
Nov 19 19:24:03 192.168.2.50 Code:
Nov 19 19:24:03 192.168.2.50 ff
Nov 19 19:24:03 192.168.2.50 8b
Nov 19 19:24:03 192.168.2.50 4d
Nov 19 19:24:03 192.168.2.50 ec
Nov 19 19:24:03 192.168.2.50 8b
Nov 19 19:24:03 192.168.2.50 41
Nov 19 19:24:03 192.168.2.50 2c
Nov 19 19:24:03 192.168.2.50 e8
Nov 19 19:24:03 192.168.2.50 bd
Nov 19 19:24:03 192.168.2.50 55
Nov 19 19:24:03 192.168.2.50 17
Nov 19 19:24:03 192.168.2.50 00
Nov 19 19:24:03 192.168.2.50 89
Nov 19 19:24:03 192.168.2.50 45
Nov 19 19:24:03 192.168.2.50 f0
Nov 19 19:24:03 192.168.2.50 89
Nov 19 19:24:03 192.168.2.50 1c
Nov 19 19:24:03 192.168.2.50 24
Nov 19 19:24:03 192.168.2.50 e8
Nov 19 19:24:03 192.168.2.50 51
Nov 19 19:24:03 192.168.2.50 be
Nov 19 19:24:03 192.168.2.50 ff
Nov 19 19:24:03 192.168.2.50 ff
Nov 19 19:24:03 192.168.2.50 eb
Nov 19 19:24:03 192.168.2.50 ad
Nov 19 19:24:03 192.168.2.50 55
Nov 19 19:24:03 192.168.2.50 89
Nov 19 19:24:03 192.168.2.50 e5
Nov 19 19:24:03 192.168.2.50 57
Nov 19 19:24:03 192.168.2.50 56
Nov 19 19:24:03 192.168.2.50 53
Nov 19 19:24:03 192.168.2.50 83
Nov 19 19:24:03 192.168.2.50 ec
Nov 19 19:24:03 192.168.2.50 08
Nov 19 19:24:03 192.168.2.50 8b
Nov 19 19:24:03 192.168.2.50 55
Nov 19 19:24:03 192.168.2.50 08
Nov 19 19:24:03 192.168.2.50 8b
Nov 19 19:24:03 192.168.2.50 82
Nov 19 19:24:03 192.168.2.50 ec
Nov 19 19:24:03 192.168.2.50 00
Nov 19 19:24:03 192.168.2.50 last message repeated 2 times
Nov 19 19:24:03 192.168.2.50 b>
Nov 19 19:24:03 192.168.2.50 38
Nov 19 19:24:03 192.168.2.50 80
Nov 19 19:24:03 192.168.2.50 b8
Nov 19 19:24:03 192.168.2.50 79
Nov 19 19:24:03 192.168.2.50 01
Nov 19 19:24:03 192.168.2.50 00
Nov 19 19:24:04 192.168.2.50 last message repeated 2 times
Nov 19 19:24:04 192.168.2.50 0f
Nov 19 19:24:04 192.168.2.50 88
Nov 19 19:24:04 192.168.2.50 8e
Nov 19 19:24:04 192.168.2.50 00
Nov 19 19:24:04 192.168.2.50 last message repeated 2 times
Nov 19 19:24:04 192.168.2.50 8b
Nov 19 19:24:04 192.168.2.50 47
Nov 19 19:24:04 192.168.2.50 2c
Nov 19 19:24:04 192.168.2.50 e8
Nov 19 19:24:04 192.168.2.50 87
Nov 19 19:24:04 192.168.2.50 55
Nov 19 19:24:04 192.168.2.50
Nov 19 19:24:04 192.168.2.50
Nov 19 19:24:04 192.168.2.50 kernel: Unable to handle kernel NULL pointer
dereference at virtual address 00000000
Nov 19 19:24:04 192.168.2.50 kernel:  printing eip:
Nov 19 19:24:04 192.168.2.50 kernel: c03df86c
Nov 19 19:24:04 192.168.2.50 kernel: *pde = 006d1001
Nov 19 19:24:04 192.168.2.50 kernel: Oops: 0000 [#1]
Nov 19 19:24:04 192.168.2.50 kernel: SMP
Nov 19 19:24:04 192.168.2.50 kernel: Modules linked in: netconsole
Nov 19 19:24:04 192.168.2.50 kernel: CPU:    3
Nov 19 19:24:04 192.168.2.50 Kernel panic - not syncing: Fatal exception in
interrupt
Nov 19 19:24:04 192.168.2.50 kernel: EIP:    0060:[<c03df86c>]    Not
tainted VLI
Nov 19 19:24:04 192.168.2.50
Nov 19 19:24:04 192.168.2.50 Rebooting in 5 seconds..


Thanks

Janos

Re: Does Linux support powering down SATA drives?

[Jeff Woods]

At 11:14 -0800 11/19/2005, Marc Perkel wrote:
>So - why isn't there more SATA support. Seems like this and SMART 
>aren't supported. What's up with that? Why is SATA harder than IDE?

I think it's a case of "newer" rather than "harder" (but that doesn't 
make it less frustrating).

--
Jeff Woods <kazrak+kernel@cesmail.net>  

Re: Does Linux support powering down SATA drives?

[Alistair John Strachan]

On Saturday 19 November 2005 20:25, Alan Cox wrote:
> On Sad, 2005-11-19 at 19:00 +0000, Alistair John Strachan wrote:
> > > SATA not yet, USB you could however.
> >
> > Or PATA, of course. I switch off two of my HDs 4 minutes after last use
> > with the commands:
> >
> > hdparm -S 48 /dev/hde
> > hdparm -S 48 /dev/hdg
> >
> > Isn't there a passthru patch in the works to let commands, such as the
> > one required for suspend, through to a SATA device?
>
> The latest kernels support command passthrough for SMART and the like
> but hdparm -S does not "switch off" anything. It may spin a drive down
> but the power consumption of 23 hours a day of "spun down" is
> significant, probably more than the hour it is powered up.

Interesting.

> Same as the problem with many household devices in standby that actually
> end up using as lot of power in their many "turned off" hours

My mistake, I was unaware of the difference between "Suspend" and (presumably) 
"Sleep". I've tried the latter without success on a Maxtor 120G here, does 
this work for anybody else (hdparm -Y)?

-- 
Cheers,
Alistair.

'No sense being pessimistic, it probably wouldn't work anyway.'
Third year Computer Science undergraduate.
1F2 55 South Clerk Street, Edinburgh, UK.

Re: Does Linux support powering down SATA drives?

[Alan Cox]

On Sad, 2005-11-19 at 19:00 +0000, Alistair John Strachan wrote:
> > SATA not yet, USB you could however.
> 
> Or PATA, of course. I switch off two of my HDs 4 minutes after last use with 
> the commands:
> 
> hdparm -S 48 /dev/hde
> hdparm -S 48 /dev/hdg
> 
> Isn't there a passthru patch in the works to let commands, such as the one 
> required for suspend, through to a SATA device?

The latest kernels support command passthrough for SMART and the like
but hdparm -S does not "switch off" anything. It may spin a drive down
but the power consumption of 23 hours a day of "spun down" is
significant, probably more than the hour it is powered up.

Same as the problem with many household devices in standby that actually
end up using as lot of power in their many "turned off" hours

Re: Does Linux support powering down SATA drives?

[Ondrej Zary]

Alistair John Strachan wrote:
> My mistake, I was unaware of the difference between "Suspend" and (presumably) 
> "Sleep". I've tried the latter without success on a Maxtor 120G here, does 
> this work for anybody else (hdparm -Y)?

I've tried hdparm -Y - only once - the drive went to sleep mode and did 
not came back, I had to use reset button.

-- 
Ondrej Zary

Re: Does Linux support powering down SATA drives?

[Marc Perkel]

Alan Cox wrote:

>On Sad, 2005-11-19 at 19:00 +0000, Alistair John Strachan wrote:
>  
>
>>>SATA not yet, USB you could however.
>>>      
>>>
>>Or PATA, of course. I switch off two of my HDs 4 minutes after last use with 
>>the commands:
>>
>>hdparm -S 48 /dev/hde
>>hdparm -S 48 /dev/hdg
>>
>>Isn't there a passthru patch in the works to let commands, such as the one 
>>required for suspend, through to a SATA device?
>>    
>>
>
>The latest kernels support command passthrough for SMART and the like
>but hdparm -S does not "switch off" anything. It may spin a drive down
>but the power consumption of 23 hours a day of "spun down" is
>significant, probably more than the hour it is powered up.
>
>Same as the problem with many household devices in standby that actually
>end up using as lot of power in their many "turned off" hours
>  
>

I didn't actually mean totally power off. Spin down would be fine with 
me. Just seems like a waste to run a drive for 24 hours that is used 
only for 10 minutes. That drive is there so if the main drive blows I 
can run down to the datacenter and move one cable and be back up again.

You know what's interesting is that I read somewhere that computers use 
as much power as 4 hoover dams can generate. And since a lit of these 
computer are running Linux just a few lines of code can create enough 
energy savings to perhaps power a small city. Kind of amazing when you 
think about it.

SATA isn't really "new" any more. I personally consider IDE to be 
obsolete. Seems to me that Linux should fully support SATA to the same 
level as IDE drives. And I'm saying that as someone who doesn't have to 
actually code it. But I will leave messages of praise and thanks in this 
mailing list if you all catch up.

Re: Does Linux support powering down SATA drives?

[Alan Cox]

On Sad, 2005-11-19 at 20:21 +0000, Alistair John Strachan wrote:
> > Same as the problem with many household devices in standby that actually
> > end up using as lot of power in their many "turned off" hours
> 
> My mistake, I was unaware of the difference between "Suspend" and (presumably) 
> "Sleep". I've tried the latter without success on a Maxtor 120G here, does 
> this work for anybody else (hdparm -Y)?

The power consumption situation is as true for suspend as sleep. Only
pulling the plug makes that difference. At the moment the SATA layer
doesn't support hot unplugging the drive which is what is really needed.

Re: Does Linux support powering down SATA drives?

[Alistair John Strachan]

On Saturday 19 November 2005 21:20, Marc Perkel wrote:
[snip]
>
> SATA isn't really "new" any more. I personally consider IDE to be
> obsolete. Seems to me that Linux should fully support SATA to the same
> level as IDE drives. And I'm saying that as someone who doesn't have to
> actually code it. But I will leave messages of praise and thanks in this
> mailing list if you all catch up.

As Alan mentions in another thread, what is needed is true hotplug support, 
which is difficult with some controllers for which we have poor (or no) 
documentation. I wouldn't hold your breath.

As for pass-thru standby/sleep commands, as long as the pass-thru patch got 
into mainline, try a very recent version of hdparm which should understand 
sending the ATA commands over SCSI (libata).

-- 
Cheers,
Alistair.

'No sense being pessimistic, it probably wouldn't work anyway.'
Third year Computer Science undergraduate.
1F2 55 South Clerk Street, Edinburgh, UK.

Re: Does Linux support powering down SATA drives?

[James Courtier-Dutton]

Marc Perkel wrote:
> Trying to save power consumption. I have a backup drive that is used 
> only once a day to back up the main drive. So - why should I run it more 
> that 10 minutes a day? What I'd like to do is keep it in an off state 
> and then at night power it on, mount it up, do the backup, unmount it, 
> and shut it down. Can I do that?
> 

Support is being added soon, I think the latest 2.6.15-rc1 supports 
passthru. I had to install an extra patch due to a bug where it used the 
same IDE device irrespective of one asking for /dev/sda or /dev/sdb
I think the fix will be in rc2.

hdparm -S60 works fine for me. It spins down, and therefore stay nices a 
cool.
hdparm -y should work, and is like an immeadiate -S
hdparm -Y is kind of dangerous at the moment. It powers down the drive, 
but will not come back unless the entire IDE bus goes through a reset. 
The feature to support that is called hotplug, but that is not 
implemented yet for SATA. So, basically, if you do hdparm -Y, you won't 
be able to wake it up again at the moment. At least hdparm -S60 is a 
start towards what you want.

James

Re: Does Linux support powering down SATA drives?

[Rogério Brito]

On Nov 19 2005, Alan Cox wrote:
> On Sad, 2005-11-19 at 09:41 -0800, Marc Perkel wrote:
> > Trying to save power consumption.
> 
> SATA not yet, USB you could however.

And what about Firewire? I see that MacOS X automatically spins down the
drive when I unmount it from a Mac.

Would it be possible to have this feature available for userspace (where
I would think would be the proper place for the policy controlling this
thing)?


Thanks for any comments, Rogério Brito.

-- 
Rogério Brito : rbrito@ime.usp.br : http://www.ime.usp.br/~rbrito
Homepage of the algorithms package : http://algorithms.berlios.de
Homepage on freshmeat:  http://freshmeat.net/projects/algorithms/

Re: Does Linux support powering down SATA drives?

[Rogério Brito]

On Nov 19 2005, Alan Cox wrote:
> It may spin a drive down but the power consumption of 23 hours a day
> of "spun down" is significant, probably more than the hour it is
> powered up.

Of course, so obvious now that you point it out.


Thanks, Rogério Brito.

-- 
Rogério Brito : rbrito@ime.usp.br : http://www.ime.usp.br/~rbrito
Homepage of the algorithms package : http://algorithms.berlios.de
Homepage on freshmeat:  http://freshmeat.net/projects/algorithms/

resyncing broken software raid 1

[Marc Perkel]

OK - I must be blind but what do you do in FC4 to resync a borken raid 1
array? It's software raid. I thought it was raidhotadd but can't get that
to work.

What am I missig?

Re: resyncing broken software raid 1

[Kyle Moffett]

On Nov 19, 2005, at 21:09:39, Marc Perkel wrote:

> OK - I must be blind but what do you do in FC4 to resync a borken  
> raid 1 array? It's software raid. I thought it was raidhotadd but  
> can't get that to work.
>

Please create a new thread (instead of responding to a message in an  
existing thread) when you want to discuss a new topic.

See "man mdadm" for more info, raidtools have been deprecated for a  
long time.

Cheers,
Kyle Moffett

Re: Does Linux support powering down SATA drives?

[asmith]

I would agree with your view on IDE becoming obsolete on hard drives, but I as 
yet, am not aware of any CD/DVD drives with a SATA interface.




On Sat, 19 Nov 2005, Alistair John Strachan wrote:

> On Saturday 19 November 2005 21:20, Marc Perkel wrote:
> [snip]
>>
>> SATA isn't really "new" any more. I personally consider IDE to be
>> obsolete. Seems to me that Linux should fully support SATA to the same
>> level as IDE drives. And I'm saying that as someone who doesn't have to
>> actually code it. But I will leave messages of praise and thanks in this
>> mailing list if you all catch up.
>
> As Alan mentions in another thread, what is needed is true hotplug support,
> which is difficult with some controllers for which we have poor (or no)
> documentation. I wouldn't hold your breath.
>
> As for pass-thru standby/sleep commands, as long as the pass-thru patch got
> into mainline, try a very recent version of hdparm which should understand
> sending the ATA commands over SCSI (libata).
>
>

Re: Does Linux support powering down SATA drives?

[Marc Perkel]

asmith@vtrl.co.uk wrote:

> I would agree with your view on IDE becoming obsolete on hard drives, 
> but I as yet, am not aware of any CD/DVD drives with a SATA interface.
>

That's something that amazes me - why aren't there anu SATA CD/DVD 
drives? Seems to me that CD/DVD could have the same benifits of using a 
thin cable and eventually getting away from the IDE interface. Or - 
since there are usually plenty of USB 2 ports, why not have a CD/DVD 
that internally connect to USB?

Another thing I've wondered about was why not build a 3 1/2 inch CD/DVD 
drive? There are plenty of the small disks available. Seems like a 
logical product to develop.

Re: Does Linux support powering down SATA drives?

[Tomasz Torcz]

[-- Attachment #1: Type: text/plain, Size: 671 bytes --]

On Sun, Nov 20, 2005 at 07:53:49AM -0800, Marc Perkel wrote:
> 
> 
> asmith@vtrl.co.uk wrote:
> 
> >I would agree with your view on IDE becoming obsolete on hard drives, 
> >but I as yet, am not aware of any CD/DVD drives with a SATA interface.
> >
> 
> That's something that amazes me - why aren't there anu SATA CD/DVD 
> drives? Seems to me that CD/DVD could have the same benifits of using a 
> thin cable and eventually getting away from the IDE interface. 

 You can rip one from an Xbox 360 :)

-- 
Tomasz Torcz                 Morality must always be based on practicality.
zdzichu@irc.-nie.spam-.pl                -- Baron Vladimir Harkonnen


[-- Attachment #2: Type: application/pgp-signature, Size: 229 bytes --]

Re: Does Linux support powering down SATA drives?

[Marc Perkel]

Tomasz Torcz wrote:

>On Sun, Nov 20, 2005 at 07:53:49AM -0800, Marc Perkel wrote:
>  
>
>>asmith@vtrl.co.uk wrote:
>>
>>    
>>
>>>I would agree with your view on IDE becoming obsolete on hard drives, 
>>>but I as yet, am not aware of any CD/DVD drives with a SATA interface.
>>>
>>>      
>>>
>>That's something that amazes me - why aren't there anu SATA CD/DVD 
>>drives? Seems to me that CD/DVD could have the same benifits of using a 
>>thin cable and eventually getting away from the IDE interface. 
>>    
>>
>
> You can rip one from an Xbox 360 :)
>
>  
>

Oh? What does the Xbox have in it?

-- 
Marc Perkel - marc@perkel.com

Spam Filter: http://www.junkemailfilter.com
    My Blog: http://marc.perkel.com

Re: Does Linux support powering down SATA drives?

[Alistair John Strachan]

On Sunday 20 November 2005 07:22, asmith@vtrl.co.uk wrote:
> I would agree with your view on IDE becoming obsolete on hard drives, but I
> as yet, am not aware of any CD/DVD drives with a SATA interface.
>

Pioneer make a SATA DVD writer.

-- 
Cheers,
Alistair.

'No sense being pessimistic, it probably wouldn't work anyway.'
Third year Computer Science undergraduate.
1F2 55 South Clerk Street, Edinburgh, UK.

Re: Does Linux support powering down SATA drives?

[Tomasz Torcz]

On Sun, Nov 20, 2005 at 08:07:07AM -0800, Marc Perkel wrote:
> Tomasz Torcz wrote:
> 
> >On Sun, Nov 20, 2005 at 07:53:49AM -0800, Marc Perkel wrote:
> > 
> >
> >>asmith@vtrl.co.uk wrote:
> >>
> >>>I would agree with your view on IDE becoming obsolete on hard drives, 
> >>>but I as yet, am not aware of any CD/DVD drives with a SATA interface.
> >>>
> >>That's something that amazes me - why aren't there anu SATA CD/DVD 
> >>drives? Seems to me that CD/DVD could have the same benifits of using a 
> >>thin cable and eventually getting away from the IDE interface. 
> >>   
> >>
> >
> >You can rip one from an Xbox 360 :)
> >
> 
> Oh? What does the Xbox have in it?

 Xbox *360* - http://www.anandtech.com/systems/showdoc.aspx?i=2610&p=4
Very thin on details, unfortunately.
 
-- 
Tomasz Torcz                 Morality must always be based on practicality.
zdzichu@irc.-nie.spam-.pl                -- Baron Vladimir Harkonnen

Re: Does Linux support powering down SATA drives?

[Pavel Machek]

On Sat 19-11-05 20:25:07, Alan Cox wrote:
> On Sad, 2005-11-19 at 19:00 +0000, Alistair John Strachan wrote:
> > > SATA not yet, USB you could however.
> > 
> > Or PATA, of course. I switch off two of my HDs 4 minutes after last use with 
> > the commands:
> > 
> > hdparm -S 48 /dev/hde
> > hdparm -S 48 /dev/hdg
> > 
> > Isn't there a passthru patch in the works to let commands, such as the one 
> > required for suspend, through to a SATA device?
> 
> The latest kernels support command passthrough for SMART and the like
> but hdparm -S does not "switch off" anything. It may spin a drive down
> but the power consumption of 23 hours a day of "spun down" is
> significant, probably more than the hour it is powered up.

Really? Harddrive does not contain AC/DC converters, so situation should be slightly
better there, no?
								Pavel
-- 
64 bytes from 195.113.31.123: icmp_seq=28 ttl=51 time=448769.1 ms         

Re: resyncing broken software raid 1

[Jens Axboe]

On Sat, Nov 19 2005, Kyle Moffett wrote:
> On Nov 19, 2005, at 21:09:39, Marc Perkel wrote:
> 
> >OK - I must be blind but what do you do in FC4 to resync a borken  
> >raid 1 array? It's software raid. I thought it was raidhotadd but  
> >can't get that to work.
> >
> 
> Please create a new thread (instead of responding to a message in an  
> existing thread) when you want to discuss a new topic.

And even better, ask in a forum appropriate to your question.
Unfortunately some people have a bad habbit of posting completely off
topic questions on lkml.

-- 
Jens Axboe

Re: Does Linux support powering down SATA drives?

[Stefan Seyfried]

Pavel Machek wrote:
> On Sat 19-11-05 20:25:07, Alan Cox wrote:

>> The latest kernels support command passthrough for SMART and the like
>> but hdparm -S does not "switch off" anything. It may spin a drive down
>> but the power consumption of 23 hours a day of "spun down" is
>> significant, probably more than the hour it is powered up.
> 
> Really? Harddrive does not contain AC/DC converters, so situation should be slightly
> better there, no?

it is 7.5W vs. 0.9W (idle vs standby or sleep) on a Barracuda 120GB
7200rpm desktop drive. Notebook drives are doing better:

0.65W vs 0.25W (idle vs standby) or
0.65W vs 0.1W  (idle vs sleep) on some Hitachi Travelstar.

Better than nothing, but pulling the plug is even better :-)
-- 
Stefan Seyfried                  \ "I didn't want to write for pay. I
QA / R&D Team Mobile Devices      \ wanted to be paid for what I write."
SUSE LINUX Products GmbH, Nürnberg \                    -- Leonard Cohen

Re: Does Linux support powering down SATA drives?

[Mark Lord]

Yes, the hdparm -y, -Y, and -S flags all work with the passthru feature set,
which is included in the 2.6.15-rc* kernels.

Typical power consumption figures, from WDC SataII drives:

Idle:  430mA@12VDC + 730mA@5VDC (about 8.75 watts)
Standby:  20mA@12VDC + 270mA@5VDC (about 1.60 watts)
Sleep: 20mA@12VDC + 250mA@5VDC (about 1.50 watts)

Those are from the WDC datasheets.

Cheers

Re: Does Linux support powering down SATA drives?

[Marc Perkel]

Mark Lord wrote:

> Yes, the hdparm -y, -Y, and -S flags all work with the passthru 
> feature set,
> which is included in the 2.6.15-rc* kernels.
>
> Typical power consumption figures, from WDC SataII drives:
>
> Idle:  430mA@12VDC + 730mA@5VDC (about 8.75 watts)
> Standby:  20mA@12VDC + 270mA@5VDC (about 1.60 watts)
> Sleep: 20mA@12VDC + 250mA@5VDC (about 1.50 watts)
>
> Those are from the WDC datasheets.
>
> Cheers


Here's maxotor specs

Power
Spinup (avg, watts) 28.7
Seek (avg, watts) 12.3
Idle (avg, watts) 7.0
Standby (avg, watts) 2.1

-- 
Marc Perkel - marc@perkel.com

Spam Filter: http://www.junkemailfilter.com
    My Blog: http://marc.perkel.com

Re: Does Linux support powering down SATA drives?

[Rob Landley]

On Sunday 20 November 2005 01:22, asmith@vtrl.co.uk wrote:
> I would agree with your view on IDE becoming obsolete on hard drives, but I
> as yet, am not aware of any CD/DVD drives with a SATA interface.

Laptops?

Rob

Re: Does Linux support powering down SATA drives?

[Jeff Garzik]

On Mon, Nov 21, 2005 at 10:21:21AM -0600, Rob Landley wrote:
> On Sunday 20 November 2005 01:22, asmith@vtrl.co.uk wrote:
> > I would agree with your view on IDE becoming obsolete on hard drives, but I
> > as yet, am not aware of any CD/DVD drives with a SATA interface.
> 
> Laptops?

Laptops will likely get S/ATAPI later rather than sooner.

But there are definitely S/ATAPI devices out there.  Most are first
generation, which means they are PATA devices with a SATA bridge (thus
they appear to the end user to be SATA).

	Jeff

Re: Does Linux support powering down SATA drives?

[Jason L Tibbitts III]

>>>>> "a" == asmith  <asmith@vtrl.co.uk> writes:

a> I would agree with your view on IDE becoming obsolete on hard
a> drives, but I as yet, am not aware of any CD/DVD drives with a SATA
a> interface.

The Plextor PX-716SA is pretty common and very nice; I have one in a
(Windows) gaming box I put together about six months ago.

 - J<

Re: Does Linux support powering down SATA drives?

[Rob Landley]

On Monday 21 November 2005 10:51, Jeff Garzik wrote:
> On Mon, Nov 21, 2005 at 10:21:21AM -0600, Rob Landley wrote:
> > On Sunday 20 November 2005 01:22, asmith@vtrl.co.uk wrote:
> > > I would agree with your view on IDE becoming obsolete on hard drives,
> > > but I as yet, am not aware of any CD/DVD drives with a SATA interface.
> >
> > Laptops?
>
> Laptops will likely get S/ATAPI later rather than sooner.

Sorry, I meant IDE isn't becoming obsolete on laptops just yet.

Eventually, sure...

Rob

Re: Does Linux support powering down SATA drives?

[Robert Hancock]

asmith@vtrl.co.uk wrote:
> I would agree with your view on IDE becoming obsolete on hard drives, 
> but I as yet, am not aware of any CD/DVD drives with a SATA interface.

They are still uncommon, but do exist. I suspect part of the reason is 
that the advantages of SATA aren't as compelling for optical drives, and 
not everybody has a machine that supports SATA.

As pointed out, the XBox 360 uses an SATA DVD-ROM drive..

-- 
Robert Hancock      Saskatoon, SK, Canada
To email, remove "nospam" from hancockr@nospamshaw.ca
Home Page: http://www.roberthancock.com/

Re: Does Linux support powering down SATA drives?

[Nicolas Mailhot]

[-- Attachment #1: Type: text/plain, Size: 1514 bytes --]

> I would agree with your view on IDE becoming obsolete on hard drives,
> but I as yet, am not aware of any CD/DVD drives with a SATA interface.

$  cat /proc/scsi/scsi
Attached devices:
Host: scsi1 Channel: 00 Id: 00 Lun: 00
  Vendor: PLEXTOR  Model: DVDR   PX-716A   Rev: 1.09
  Type:   CD-ROM                           ANSI SCSI revision: 05
Host: scsi4 Channel: 00 Id: 00 Lun: 00
  Vendor: ATA      Model: Maxtor 6L300S0   Rev: BANC
  Type:   Direct-Access                    ANSI SCSI revision: 05
Host: scsi5 Channel: 00 Id: 00 Lun: 00
  Vendor: ATA      Model: Maxtor 6L300S0   Rev: BANC
  Type:   Direct-Access                    ANSI SCSI revision: 05
$ /sbin/lspci | grep -i scsi
$ /sbin/lspci | grep -i ata
00:07.0 IDE interface: nVidia Corporation CK804 Serial ATA Controller
(rev f3)
00:08.0 IDE interface: nVidia Corporation CK804 Serial ATA Controller
(rev f3)
01:09.0 Mass storage controller: Silicon Image, Inc. SiI 3114
[SATALink/SATARaid] Serial ATA Controller (rev 02)

And BTW,

#  /sbin/hdparm -M 128 /dev/hda

/dev/hda:
 setting acoustic management to 128
 acoustic     =  0 (128=quiet ... 254=fast)

(failure but I don't care the drive is old and already in AAM mode)

#  /sbin/hdparm -M 128 /dev/sda

/dev/sda:
 setting acoustic management to 128
 HDIO_GET_ACOUSTIC failed: Inappropriate ioctl for device

(This drive however is not and needs it dearly. Plus it's the Linux
drive, so it's used most of the time)

Regards,

-- 
Nicolas Mailhot

[-- Attachment #2: Ceci est une partie de message numériquement signée --]
[-- Type: application/pgp-signature, Size: 197 bytes --]

Re: Does Linux support powering down SATA drives?

[Douglas Gilbert]

Marc Perkel wrote:
> Trying to save power consumption. I have a backup drive that is used
> only once a day to back up the main drive. So - why should I run it more
> that 10 minutes a day? What I'd like to do is keep it in an off state
> and then at night power it on, mount it up, do the backup, unmount it,
> and shut it down. Can I do that?

Yes, from lk 2.6.14 onwards.

An implementation (based on SAT:
http://www.t10.org/ftp/t10/drafts/sat/sat-r07.pdf) of
the START STOP UNIT SCSI command went into libata in
lk 2.6.14 . Hence you can use SCSI tools such as
sg_start (sg3_utils) or sdparm (sdparm) to place a
SATA disk in standby mode. For example:
"sg_start 0 /dev/sda" and "sdparm --command=stop /dev/sda"
are equivalent to "hdparm -y". The next command sent to that
disk will cause it to spin up again.

The power state machines of SCSI disks, ATA disks
and CD/DVD drives overlap but are not the same.
SATA and SAS transports add some wrinkles, see:
http://www.torque.net/sg/power.html
Spinning down disks with no mounted file systems makes
sense. However repeatedly spinning down a disk that is
periodically (e.g. 30 seconds later) accessed may shorten
its life.

In lk 2.6.15-rc1 implementations of the ATA COMMAND
PASS THROUGH (12 and 16 byte) SCSI commands went into
libata. These are defined in SAT (reference above).
libata also translates the HDIO_DRIVE_CMD and
HDIO_DRIVE_TASK ioctls into those pass through commands.
Recent versions of hdparm (I've tried 6.1 and 6.3) work as
expected. Hence "hdparm -y /dev/sda" puts the libata-connected
SATA disk /dev/sda into standby mode.
smartmontools also works for libata-connected SATA disks
in lk 2.6.15-rc1 .

The picture is not as bright for external USB and
1394 enclosures of ATA disks. They need to either
support the START STOP UNIT SCSI command or the
SAT pass through commands. I have not seen the
former and won't hold my breath waiting for the
latter.

Doug Gilbert