I can't get "hdparm -S" to work at all.
Using "hdparm -S 1" should set the timeout to
5 seconds, but the drives stay active/idle all the
time. When I set to standby manually, the drives
stay on standby for days, and start up fine when
they are used. I know this because I logged status
every 30 minutes using crontab.
They just don't standby automatically.
I'm using 5 identical drives (see below for drive
info), with software RAID.
Any suggestions? System details follow.
# uname -a
Linux narf 2.6.22-gentoo-r5 #1 Sat Sep 1 13:35:06 CDT 2007 i686
Intel(R) Celeron(R) D CPU 420 @ 1.60GHz GenuineIntel GNU/Linux
# hdparm -S 1 /dev/sda
setting standby to 1 (5 seconds)
# hdparm -I /dev/sda
/dev/sda:
ATA device, with non-removable media
Model Number: WDC WD5000ABYS-01TNA0
Serial Number: WD-WCAPW2171884
Firmware Revision: 12.01C01
Standards:
Supported: 7 6 5 4
Likely used: 7
Configuration:
Logical max current
cylinders 16383 16383
heads 16 16
sectors/track 63 63
--
CHS current addressable sectors: 16514064
LBA user addressable sectors: 268435455
LBA48 user addressable sectors: 976773168
device size with M = 1024*1024: 476940 MBytes
device size with M = 1000*1000: 500107 MBytes (500 GB)
Capabilities:
LBA, IORDY(can be disabled)
Queue depth: 32
Standby timer values: spec'd by Standard, with device specific minimum
R/W multiple sector transfer: Max = 16 Current = 16
Recommended acoustic management value: 128, current value: 254
DMA: mdma0 mdma1 mdma2 udma0 udma1 udma2 udma3 udma4 udma5 *udma6
Cycle time: min=120ns recommended=120ns
PIO: pio0 pio1 pio2 pio3 pio4
Cycle time: no flow control=120ns IORDY flow control=120ns
Commands/features:
Enabled Supported:
* SMART feature set
Security Mode feature set
* Power Management feature set
* Write cache
* Look-ahead
* Host Protected Area feature set
* WRITE_BUFFER command
* READ_BUFFER command
* NOP cmd
* DOWNLOAD_MICROCODE
Power-Up In Standby feature set
* SET_FEATURES required to spinup after power up
SET_MAX security extension
* Automatic Acoustic Management feature set
* 48-bit Address feature set
* Device Configuration Overlay feature set
* Mandatory FLUSH_CACHE
* FLUSH_CACHE_EXT
* SMART error logging
* SMART self-test
* General Purpose Logging feature set
* WRITE_{DMA|MULTIPLE}_FUA_EXT
* 64-bit World wide name
* Segmented DOWNLOAD_MICROCODE
* SATA-I signaling speed (1.5Gb/s)
* SATA-II signaling speed (3.0Gb/s)
* Native Command Queueing (NCQ)
* Phy event counters
DMA Setup Auto-Activate optimization
* Software settings preservation
* SMART Command Transport (SCT) feature set
* SCT Long Sector Access (AC1)
* SCT LBA Segment Access (AC2)
* SCT Error Recovery Control (AC3)
* SCT Features Control (AC4)
* SCT Data Tables (AC5)
unknown 206[12]
unknown 206[13]
Security:
Master password revision code = 65534
supported
not enabled
not locked
frozen
not expired: security count
not supported: enhanced erase
134min for SECURITY ERASE UNIT.
Checksum: correct
Mark Weber wrote:
> I can't get "hdparm -S" to work at all.
>
> Using "hdparm -S 1" should set the timeout to
> 5 seconds, but the drives stay active/idle all the
> time. When I set to standby manually, the drives
> stay on standby for days, and start up fine when
> they are used. I know this because I logged status
> every 30 minutes using crontab.
>
> They just don't standby automatically.
>
> I'm using 5 identical drives (see below for drive
> info), with software RAID.
I don't have Raptors here anymore to try it with,
but "hdparm -S1 /dev/sd?" did work fine with them
when last tried.
What do you get when you try this:
hdparm -S1 -C /dev/sda;sleep 6; hdparm -C /dev/sda
??
On 10/11/07, Mark Lord <[email protected]> wrote:
>
> What do you get when you try this:
> hdparm -S1 -C /dev/sda;sleep 6; hdparm -C /dev/sda
> ??
Thanks for taking an interest.
I ran the commands and got the result shown
below. Anything else you want me to try?
# hdparm -S1 -C /dev/sda; sleep 6; hdparm -C /dev/sda
/dev/sda:
setting standby to 1 (5 seconds)
drive state is: active/idle
/dev/sda:
drive state is: active/idle
Should I give sdparm a shot?
Mark Weber wrote:
>
> # hdparm -S1 -C /dev/sda; sleep 6; hdparm -C /dev/sda
>
> /dev/sda:
> setting standby to 1 (5 seconds)
> drive state is: active/idle
>
> /dev/sda:
> drive state is: active/idle
That's interesting.
So, either something is regularly accessing/polling the drive,
or it just doesn't work with the standby timer.
Are there any interesting kernel messages being generated
during execution of those commands?
-ml
On 10/12/07, Mark Lord <[email protected]> wrote:
>
> That's interesting.
>
> So, either something is regularly accessing/polling the drive,
> or it just doesn't work with the standby timer.
>
> Are there any interesting kernel messages being generated
> during execution of those commands?
No messages to /var/log/messages as a result of those
commands; I ran several times and did a sync too, just
in case.
I very much doubt that something is regularly accessing
the drives because I have a script that runs every 1/2 hour
to check and log status. The drives stay "standby" for
hours (even days) at a time unless I specifically access
something on the RAID array. Then, they stay "active/idle"
until I manually set them to "standby".
Anything else you'd like me to try?
Mark Weber wrote:
>
> Anything else you'd like me to try?
Yes: a stock kernel, please. Preferably 2.6.23.
Thanks.
Mark Weber wrote:
> On 10/12/07, Mark Lord <[email protected]> wrote:
>> That's interesting.
>>
>> So, either something is regularly accessing/polling the drive,
>> or it just doesn't work with the standby timer.
>>
>> Are there any interesting kernel messages being generated
>> during execution of those commands?
>
> No messages to /var/log/messages as a result of those
> commands; I ran several times and did a sync too, just
> in case.
>
> I very much doubt that something is regularly accessing
> the drives because I have a script that runs every 1/2 hour
> to check and log status. The drives stay "standby" for
> hours (even days) at a time unless I specifically access
> something on the RAID array. Then, they stay "active/idle"
> until I manually set them to "standby".
>
> Anything else you'd like me to try?
Some things to check:
* Run "hdparm -I" on your drive. In the "Capabilities" section there is
a line "Standby timer values", for some drives this mentions a device
specific minimum. I know some drives that ignore any setting below 60
seconds.
* I also know of quite a number of drives where hdparm -B settings
override the -S settings, even if you set the -S settings after the
hdparm -B settings. You could try combinations with various values of
hdparm -B, especially 1 and 255.
Cheers,
Bart
On 10/14/07, Bart Samwel <[email protected]> wrote:
>
> Some things to check:
>
> * Run "hdparm -I" on your drive. In the "Capabilities" section there is
> a line "Standby timer values", for some drives this mentions a device
> specific minimum. I know some drives that ignore any setting below 60
> seconds.
>
> * I also know of quite a number of drives where hdparm -B settings
> override the -S settings, even if you set the -S settings after the
> hdparm -B settings. You could try combinations with various values of
> hdparm -B, especially 1 and 255.
Thanks for the suggestions.
The -I command prints out a bunch of stuff including:
Standby timer values: spec'd by Standard, with device specific minimum
I tried setting -B to 1 and and then set -S to 5 minutes.
Also, -B 255 and then set -S to 5 minutes.
No luck with either. These drives want to keep running.
One thing of possible interest: The -B command printed
the following message:
/dev/sda:
setting Advanced Power Management level to 0x01 (1)
HDIO_DRIVE_CMD failed: Input/output error
I would guess that the first line came out just before
hdparm tried to do the set, and the second line indicates
that the set failed.
Perhaps -S is failing too, just without the diagnostic?
All of the above using 2.6.22-gentoo-r5
I have not yet tried 2.6.23 as suggested by Mark Lord.
---
"And he puzzled three hours, `till his puzzler was sore."
from "How the Grinch Stole Christmas"
Mark Weber wrote:
> On 10/14/07, Bart Samwel <[email protected]> wrote:
>> Some things to check:
>>
>> * Run "hdparm -I" on your drive. In the "Capabilities" section there is
>> a line "Standby timer values", for some drives this mentions a device
>> specific minimum. I know some drives that ignore any setting below 60
>> seconds.
>>
>> * I also know of quite a number of drives where hdparm -B settings
>> override the -S settings, even if you set the -S settings after the
>> hdparm -B settings. You could try combinations with various values of
>> hdparm -B, especially 1 and 255.
>
> Thanks for the suggestions.
>
> The -I command prints out a bunch of stuff including:
> Standby timer values: spec'd by Standard, with device specific minimum
Ahhh. Spec'd by standard means that each -S unit is worth 5 seconds (for
values up to 240 = 20 minutes), and the second part means that there is
a minimum (which is not specified in this report, unfortunately).
Perhaps you can get a hold of the full drive manual, the exact minimum
value is probably specified there.
> I tried setting -B to 1 and and then set -S to 5 minutes.
> Also, -B 255 and then set -S to 5 minutes.
> No luck with either. These drives want to keep running.
Just to be sure: you did use -S 60 to get 5 minutes, right?
> One thing of possible interest: The -B command printed
> the following message:
>
> /dev/sda:
> setting Advanced Power Management level to 0x01 (1)
> HDIO_DRIVE_CMD failed: Input/output error
>
> I would guess that the first line came out just before
> hdparm tried to do the set, and the second line indicates
> that the set failed.
Yes, that seems correct. Nothing too weird there: it simply seems that
the drive doesn't support the power management knob. (AFAIK you should
be able to confirm this using the feature sets listed in the output of
hdparm -I.)
> Perhaps -S is failing too, just without the diagnostic?
Perhaps, but I'd expect it to print a diagnostic if it fails. I do seem
to remember that (at least for some drives that I've seen) there isn't a
diagnostic if you go below the device specific minimum, the value is
simply ignored.
Cheers,
Bart
On 10/14/07, Bart Samwel <[email protected]> wrote:
>
> Just to be sure: you did use -S 60 to get 5 minutes, right?
Yes. And hdparm is kind enough to print:
/dev/sda:
setting standby to 60 (5 minutes)
Here's a bizarre sequence which I just noticed:
[extraneous blank lines removed for clarity]
>> hdparm -C /dev/sd[abcde]
/dev/sda: drive state is: standby
/dev/sdb: drive state is: standby
/dev/sdc: drive state is: standby
/dev/sdd: drive state is: standby
/dev/sde: drive state is: standby
>> hdparm -S 60 /dev/sda
/dev/sda: setting standby to 60 (5 minutes)
>> hdparm -C /dev/sd[abcde]
/dev/sda: drive state is: active/idle
/dev/sdb: drive state is: active/idle
/dev/sdc: drive state is: standby
/dev/sdd: drive state is: standby
/dev/sde: drive state is: standby
Note that the -S 60 on /dev/sda affected
/dev/sdb too! This is repeatable.
I have these drives as RAID5 (software RAID).
I don't know if that has anything to do with the
failure of -S or not. Don't know if hdparm bypasses
the RAID or not.
1. How are you forcing the drives into standby?
2. Have you reproduced this with a stock kernel.org kernel yet?
One possibility is that these drives may be the kind that
generate a "spurious" interrupt when they spin-down via the timer,
and perhaps libata is "processing" that interrupt and sending
additional command(s) that then wake the drive up again immediately.
To rule this out, you could try using drivers/ide for a moment or two,
and see if the same problem persists with those drives.
You could also try dumping /proc/interrupts in conjunction with "hdparm -S1",
and we can compare that with a "known good" system.
Something like this:
hdparm -B255 /dev/sda
hdparm -S1 /dev/sda
cat /proc/interrupts
sleep 6
cat /proc/interrupts
Cheers
On 10/14/07, Mark Lord <[email protected]> wrote:
> 1. How are you forcing the drives into standby?
hdparm -y
> 2. Have you reproduced this with a stock kernel.org kernel yet?
No; maybe later this week.
> One possibility is that these drives may be the kind that
> generate a "spurious" interrupt when they spin-down via the timer,
> and perhaps libata is "processing" that interrupt and sending
> additional command(s) that then wake the drive up again immediately.
>
> To rule this out, you could try using drivers/ide for a moment or two,
> and see if the same problem persists with those drives.
>
> You could also try dumping /proc/interrupts in conjunction with "hdparm -S1",
> and we can compare that with a "known good" system.
>
> Something like this:
>
> hdparm -B255 /dev/sda
> hdparm -S1 /dev/sda
> cat /proc/interrupts
> sleep 6
> cat /proc/interrupts
Here's the log for the second suggestion:
narf ~ # hdparm -B255 /dev/sda
/dev/sda:
setting Advanced Power Management level to disabled
HDIO_DRIVE_CMD failed: Input/output error
narf ~ # hdparm -S1 /dev/sda
/dev/sda:
setting standby to 1 (5 seconds)
narf ~ # cat /proc/interrupts
CPU0
0: 2268 IO-APIC-edge timer
1: 2 IO-APIC-edge i8042
8: 5 IO-APIC-edge rtc
9: 0 IO-APIC-fasteoi acpi
12: 4 IO-APIC-edge i8042
14: 206827 IO-APIC-edge ide0
16: 350813 IO-APIC-fasteoi sata_promise, uhci_hcd:usb5
17: 39596029 IO-APIC-fasteoi eth0
18: 0 IO-APIC-fasteoi uhci_hcd:usb4
19: 728947 IO-APIC-fasteoi libata, uhci_hcd:usb3
20: 0 IO-APIC-fasteoi ehci_hcd:usb1, uhci_hcd:usb2
NMI: 0
LOC: 2736793
ERR: 0
MIS: 0
narf ~ # sleep 6
narf ~ # cat /proc/interrupts
CPU0
0: 2268 IO-APIC-edge timer
1: 2 IO-APIC-edge i8042
8: 5 IO-APIC-edge rtc
9: 0 IO-APIC-fasteoi acpi
12: 4 IO-APIC-edge i8042
14: 206828 IO-APIC-edge ide0
16: 350813 IO-APIC-fasteoi sata_promise, uhci_hcd:usb5
17: 39596069 IO-APIC-fasteoi eth0
18: 0 IO-APIC-fasteoi uhci_hcd:usb4
19: 728947 IO-APIC-fasteoi libata, uhci_hcd:usb3
20: 0 IO-APIC-fasteoi ehci_hcd:usb1, uhci_hcd:usb2
NMI: 0
LOC: 2736881
ERR: 0
MIS: 0