Re: SATA disc spindown

2011-11-28 Thread Ramon Hofer 
On Sun, 27 Nov 2011 16:55:06 +0100, Lou wrote:

 Please check the wdc link [1] I gave you again, WD20EARS is on the list.
 The wiki article can't keep up all the time with newer green series
 coming out.

Ups, my fault.
Thank you very much!


-- 
To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org 
with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Archive: http://lists.debian.org/javv1q$kc1$1...@dough.gmane.org

Re: SATA disc spindown

2011-11-27 Thread Ramon Hofer 
Thanks alot for all the infos. Very helpful!
I'll try that this evening :-)


On Sat, 26 Nov 2011 15:57:03 +0100, Lou wrote:

 Maybe you can choose a solution where you unmount the array before you
 put them to sleep, just to be on the save side? I never used this, since
 I prefer to shutdown a server on inactivity, and wake it up again using
 WOL or acpi wakeup.

I'd like to shut down the server too. But this seemed too complicated for 
a noob like me ;-)
On the server there's a MythBackend, Logitech Media Server, content for 
XBMC and apache. All of them need to be checked so that the server 
doesn't shut when it shouldn't.
It seemed that putting the unused discs to sleep would be easier.

I don't know if I can check whether the raid is in use or not because it 
serves content for XBMC, Logitech Media Server and apache.


Thanks again

Ramon


-- 
To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org 
with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Archive: http://lists.debian.org/jatgoe$7sv$1...@dough.gmane.org

Re: SATA disc spindown

2011-11-27 Thread Ramon Hofer 
On Fri, 25 Nov 2011 16:39:16 +0100, Lou wrote:

 On Fri, 25 Nov 2011 15:14:31 +, Ramon Hofer wrote:
 
 
 Maybe this is important:
 I use Squeeze, 2.6.32-5-amd64, hdparm v9.32, the drive sdc is a WDC
 WD20EARS-00MVWB0 and the mainboard an ASUS P7P55D.



 It's very good you mention this - the WD Green series is a pain in the
 ass for linux users ... please have a look at the following two postings
 about problems with the green series operating under linux:
 
 http://wiki.ubuntuusers.de/WD_IntelliPark

Sorry, one more question to the wdidle3:
On the ubuntuusers page and the WD page they write, that it should only 
be used for WD1000FYPS-01ZKB0, WD7500AYPS-01ZKB0 and WD7501AYPS-01ZKB0.
But I have a different drive. So I wondered if they just didn't update 
their homepage and mean all of the Caviar Green series?


Cheers


-- 
To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org 
with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Archive: http://lists.debian.org/jathq9$7sv$2...@dough.gmane.org

Re: SATA disc spindown

2011-11-27 Thread Lou 

Hello Ramon,

Please check the wdc link [1] I gave you again, WD20EARS is on the list. 
The wiki article can't keep up all the time with newer green series 
coming out.


[1] http://wdc.custhelp.com/app/answers/detail/a_id/5357

Regards

Lou

Ramon Hofer wrote:


Sorry, one more question to the wdidle3:
On the ubuntuusers page and the WD page they write, that it should only
be used for WD1000FYPS-01ZKB0, WD7500AYPS-01ZKB0 and WD7501AYPS-01ZKB0.
But I have a different drive. So I wondered if they just didn't update
their homepage and mean all of the Caviar Green series?




--
To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org 
with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org

Archive: http://lists.debian.org/blu0-smtp3078098d6e755a8bb05d8...@phx.gbl

Re: SATA disc spindown

2011-11-26 Thread Ramon Hofer 
On Fri, 25 Nov 2011 16:39:16 +0100, Lou wrote:

 On Fri, 25 Nov 2011 15:14:31 +, Ramon Hofer wrote:
 
 
 Maybe this is important:
 I use Squeeze, 2.6.32-5-amd64, hdparm v9.32, the drive sdc is a WDC
 WD20EARS-00MVWB0 and the mainboard an ASUS P7P55D.



 It's very good you mention this - the WD Green series is a pain in the
 ass for linux users ... please have a look at the following two postings
 about problems with the green series operating under linux:
 
 http://wiki.ubuntuusers.de/WD_IntelliPark
 http://www.gossamer-threads.com/lists/mythtv/users/479998#479998
 
 In order to avoid high load cycle counts it is highly recommended to
 disable intellipark using WDIDLE3.exe, after that your load cycle counts
 should not raise anymore and the traditional ata commands like hdparm -S
 /dev/device 6 should work again, thus switch off your green drive
 after 30 seconds (int is a multiple of 5 accdoring to the hdparm
 manpage)
 
 Please take care to only use capital -S parameter, not -s ...
 otherwise it could mess up the device firmware.

Thanks alot! Great infos!!!
I will try to fix the problem tomorrow.

Wasn't able to have a closer look at the links but it seems as if I had 
to use Windows to run WDIDLE3.exe to change a firmware setting. I assume 
I can do this on an other PC and then move the disc back to the server?


Thanks you again.

Do you have any advices about the Samsung drives in the mdadm raid? Can I 
put them to sleep normally or do I have to tell mdadm that it has to spin 
down the drives?


Cheers
Ramon


-- 
To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org 
with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Archive: http://lists.debian.org/jaqiak$7i6$1...@dough.gmane.org

Re: Re: SATA disc spindown

2011-11-26 Thread Lou 

Ramon Hofer wrote:



Wasn't able to have a closer look at the links but it seems as if I had
to use Windows to run WDIDLE3.exe to change a firmware setting. I assume
I can do this on an other PC and then move the disc back to the server?



Actually this is an MS-DOS binary, which might complicate things for you 
if you no longer have a working dos environment like a bootable floppy 
or a stick installation. This is why I posted the wiki link: the article 
explains how to create a bootable freedos image with the exe included. 
Since I still had a working MS-DOS 6.22 floppy I used that one, but the 
wiki howto (Verwendung von FreeDOS) looks straightforward to me as well.


The binary archive can be downloaded from the wd support site [1] there 
is a text file inside that explains the update procedure.


[1] http://wdc.custhelp.com/app/answers/detail/a_id/5357



Do you have any advices about the Samsung drives in the mdadm raid? Can I
put them to sleep normally or do I have to tell mdadm that it has to spin
down the drives?



This seems to be a controversial topic - some people report no problems 
with md raid standby, some disagree because mdraid does not handle it 
well should one array member sleep and another member doesn't.


Maybe you can choose a solution where you unmount the array before you 
put them to sleep, just to be on the save side? I never used this, since 
I prefer to shutdown a server on inactivity, and wake it up again using 
WOL or acpi wakeup.


Good Luck!

Lou


--
To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org 
with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org

Archive: http://lists.debian.org/blu0-smtp33ba81c6d3bef2b8f7ecb2d8...@phx.gbl

Re: SATA disc spindown

2011-11-25 Thread Ramon Hofer 
On Fri, 25 Nov 2011 11:49:42 +, Ramon Hofer wrote:

 Therefore I uncommented the spindown_time line and set the value to 241
 which should spin down all the drives after 30 minutes.

To speed up testing I have set spindown_tim = 1 but the drives still 
stays active...


-- 
To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org 
with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Archive: http://lists.debian.org/jao093$a08$2...@dough.gmane.org

Re: SATA disc spindown

2011-11-25 Thread Lou 

Hallo Ramon,

On Fri, 25 Nov 2011 11:49:42 +, Ramon Hofer wrote:

Therefore I uncommented the spindown_time line and set the value to
241 which should spin down all the drives after 30 minutes.

To speed up testing I have set spindown_tim = 1 but the drives still
stays active...

Can I check the access to the discs from a log?
Or maybe I should create a sudo hdparm -C /dev/sd? log which checks 
the drive states every 15 minutes or so?


Maybe it's just the hddtemp daemon with the temperature polls keeping 
your data drive(s) awake? To find out more exactly I suggest you monitor 
r/w access in syslog:


#this will put a comment about read/write access into your syslog:
echo 1  /proc/sys/vm/block_dump

#live monitoring syslog:
tail -f /var/log/syslog

#switch off syslog r/w comments:
echo 1  /proc/sys/vm/block_dump

Hopefully this will give you the necessary hints

Cheers!


--
To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org 
with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org

Archive: http://lists.debian.org/blu0-smtp17874d40ed4d96c6f48a45ed8...@phx.gbl

Re: SATA disc spindown

2011-11-25 Thread Ramon Hofer 
On Fri, 25 Nov 2011 13:26:13 +0100, Lou wrote:

 Hallo Ramon,

Thanks for your reply Lou!


 On Fri, 25 Nov 2011 11:49:42 +, Ramon Hofer wrote:
 
  Therefore I uncommented the spindown_time line and set the value to
  241 which should spin down all the drives after 30 minutes.
 
  To speed up testing I have set spindown_tim = 1 but the drives still
  stays active...
 
  Can I check the access to the discs from a log? Or maybe I should
  create a sudo hdparm -C /dev/sd? log which checks the drive states
  every 15 minutes or so?
 
 Maybe it's just the hddtemp daemon with the temperature polls keeping
 your data drive(s) awake? To find out more exactly I suggest you monitor
 r/w access in syslog:

I have setup hddtemp to do nothing automatically (set interval to 0 and 
don't start as daemon). So this shouldn't be the problem.


 #this will put a comment about read/write access into your syslog: echo
 1  /proc/sys/vm/block_dump
 
 #live monitoring syslog:
 tail -f /var/log/syslog
 
 #switch off syslog r/w comments:
 echo 1  /proc/sys/vm/block_dump
 
 Hopefully this will give you the necessary hints

This is of great help!
Thanks alot!!


Btw: As you didn't write anything else hdparm should have no problem 
spinning down sata drives?

And do I have to restart the system or is it enough the do a /etc/init.d/
hdparm restart?

Do you know if I can check the hdparm setting somewhere? Like hdparm -C 
to show me the spindown setting?


Cheers


-- 
To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org 
with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Archive: http://lists.debian.org/jao4l0$a08$3...@dough.gmane.org

Re: SATA disc spindown

2011-11-25 Thread Ramon Hofer 
On Fri, 25 Nov 2011 13:19:28 +, Ramon Hofer wrote:

 On Fri, 25 Nov 2011 13:26:13 +0100, Lou wrote:
 
 Hallo Ramon,
 
 Thanks for your reply Lou!
 
 
 On Fri, 25 Nov 2011 11:49:42 +, Ramon Hofer wrote:
 
  Therefore I uncommented the spindown_time line and set the value to
  241 which should spin down all the drives after 30 minutes.
 
  To speed up testing I have set spindown_tim = 1 but the drives still
  stays active...
 
  Can I check the access to the discs from a log? Or maybe I should
  create a sudo hdparm -C /dev/sd? log which checks the drive states
  every 15 minutes or so?
 
 Maybe it's just the hddtemp daemon with the temperature polls keeping
 your data drive(s) awake? To find out more exactly I suggest you
 monitor r/w access in syslog:
 
 I have setup hddtemp to do nothing automatically (set interval to 0 and
 don't start as daemon). So this shouldn't be the problem.
 
 
 #this will put a comment about read/write access into your syslog: echo
 1  /proc/sys/vm/block_dump
 
 #live monitoring syslog:
 tail -f /var/log/syslog
 
 #switch off syslog r/w comments:
 echo 1  /proc/sys/vm/block_dump

When I turn logging on, I only get lots of messages from the system drive 
and not from any other. So it should work?


 And do I have to restart the system or is it enough the do a
 /etc/init.d/ hdparm restart?

When I only have the spindown_time option in hdparm.conf and I do /etc/
init.d/hdparm restart then it tells me: Setting parameters of disc: 
(none).

But when I add 
/dev/sdc {
   spindown_time = 1
}
I get Setting parameters of disc:  /dev/sdc.

The output from last reboot of grep sdc /var/log/syslog is
Nov 25 14:26:19 media-server kernel: [1.776301] sd 1:0:0:0: [sdc] 
3907029168 512-byte logical blocks: (2.00 TB/1.81 TiB)
Nov 25 14:26:19 media-server kernel: [1.776400] sd 1:0:0:0: [sdc] 
Write Protect is off
Nov 25 14:26:19 media-server kernel: [1.776402] sd 1:0:0:0: [sdc] 
Mode Sense: 00 3a 00 00
Nov 25 14:26:19 media-server kernel: [1.776425] sd 1:0:0:0: [sdc] 
Write cache: enabled, read cache: enabled, doesn't support DPO or FUA
Nov 25 14:26:19 media-server kernel: [1.776546]  sdc:
Nov 25 14:26:19 media-server kernel: [2.178369]  sdc1
Nov 25 14:26:19 media-server kernel: [2.200458] sd 1:0:0:0: [sdc] 
Attached SCSI disk

The DPO and FUA things doesn't seem to have anything to do with that?


-- 
To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org 
with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Archive: http://lists.debian.org/jao5s4$a08$4...@dough.gmane.org

Re: SATA disc spindown

2011-11-25 Thread Lou 

On -10.01.-28163 20:59, Ramon Hofer wrote:


Btw: As you didn't write anything else hdparm should have no problem
spinning down sata drives?

Do you know if I can check the hdparm setting somewhere? Like hdparm -C
to show me the spindown setting?


It depends on the drive - there is hdparm -I /dev/device that offers a 
capability and a command feature section. From the looks at my output

(where hdparm -S int works alright) there is this option:

   *Power Management feature set

Maybe this is mandatory for hdparm to work with your hardware, I'm not sure.


And do I have to restart the system or is it enough the do a /etc/init.d/
hdparm restart?



Nope, hdparm -S value /dev/drive should have immediate effect, but 
it could be different for the conf file in /etc - I hope somebody else 
can help you there, since I never used that file, but wrote a 
shellscript to set hdparm -S int upon boot.



When I turn logging on, I only get lots of messages from the system drive
and not from any other. So it should work?


At least that's a good sign for drive access, another way to check it is 
to unmount the data drive(s) completely, then set sleep time to a 
minimum with hdparm -S int drive , then wait while monitoring syslog :-)


regards

Lou


--
To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org 
with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org

Archive: http://lists.debian.org/blu0-smtp460a4e58174d9c4ffbf2b5ad8...@phx.gbl

Re: SATA disc spindown

2011-11-25 Thread Ramon Hofer 
On Fri, 25 Nov 2011 14:58:33 +0100, Lou wrote:

 On -10.01.-28163 20:59, Ramon Hofer wrote:
 
 Btw: As you didn't write anything else hdparm should have no problem
 spinning down sata drives?

 Do you know if I can check the hdparm setting somewhere? Like hdparm -C
 to show me the spindown setting?
 
 It depends on the drive - there is hdparm -I /dev/device that offers a
 capability and a command feature section. From the looks at my output
 (where hdparm -S int works alright) there is this option:
 
 *Power Management feature set
 
 Maybe this is mandatory for hdparm to work with your hardware, I'm not
 sure.

When I set hdparm -S it tells me it sets the option and what the number 
means.
hdparm -I says the Power Management feature is set (btw it is still set 
when I set hdparm -S 0).
So far so good.

But even when I unmount the drive it doesn't spin down, at least there's 
nothing in the syslog and hdparm -C says it's still active.

I can unmount the disc, set it to sleep with -Y and when I mount it again 
syslog says ata2.00: waking up from sleep

Nov 25 16:00:01 media-server kernel: [ 1055.621588] ata2.00: exception 
Emask 0x0 SAct 0x0 SErr 0x0 action 0x6
Nov 25 16:00:01 media-server kernel: [ 1055.621664] ata2.00: waking up 
from sleep
Nov 25 16:00:01 media-server kernel: [ 1055.621738] ata2.00: hard 
resetting link
Nov 25 16:00:01 media-server kernel: [ 1055.938034] ata2.01: hard 
resetting link
Nov 25 16:00:02 media-server kernel: [ 1056.414344] ata2.00: SATA link up 
3.0 Gbps (SStatus 123 SControl 300)
Nov 25 16:00:02 media-server kernel: [ 1056.414360] ata2.01: SATA link up 
3.0 Gbps (SStatus 123 SControl 300)
Nov 25 16:00:02 media-server kernel: [ 1056.438445] ata2.00: configured 
for UDMA/133
Nov 25 16:00:02 media-server kernel: [ 1056.446699] ata2.01: configured 
for UDMA/133
Nov 25 16:00:02 media-server kernel: [ 1056.446711] ata2: EH complete

Seems ok. But nothing is syslogged when I mount it after setting 
hdparm -S 1.

Any other ideas?


Maybe this is important:
I use Squeeze, 2.6.32-5-amd64, hdparm v9.32, the drive sdc is a WDC 
WD20EARS-00MVWB0 and the mainboard an ASUS P7P55D.

There are four SAMSUNG HD154UI discs in a mdadm raid5 array but I think 
they more complicated to setup because of mdadm. The idea was to send 
them to sleep too but because of mdadm it seems to be a better idea to 
look at them later...


-- 
To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org 
with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Archive: http://lists.debian.org/jaob2f$a08$5...@dough.gmane.org

Re: SATA disc spindown

2011-11-25 Thread Lou 

On Fri, 25 Nov 2011 15:14:31 +, Ramon Hofer wrote:



Maybe this is important:
I use Squeeze, 2.6.32-5-amd64, hdparm v9.32, the drive sdc is a WDC
WD20EARS-00MVWB0 and the mainboard an ASUS P7P55D.




It's very good you mention this - the WD Green series is a pain in the 
ass for linux users ... please have a look at the following two postings 
about problems with the green series operating under linux:


http://wiki.ubuntuusers.de/WD_IntelliPark
http://www.gossamer-threads.com/lists/mythtv/users/479998#479998

In order to avoid high load cycle counts it is highly recommended to 
disable intellipark using WDIDLE3.exe, after that your load cycle counts 
should not raise anymore and the traditional ata commands like hdparm -S 
/dev/device 6 should work again, thus switch off your green drive 
after 30 seconds (int is a multiple of 5 accdoring to the hdparm manpage)


Please take care to only use capital -S parameter, not -s ... 
otherwise it could mess up the device firmware.


Regards

Lou


--
To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org 
with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org

Archive: http://lists.debian.org/blu0-smtp301232f102d76b5569fea47d8...@phx.gbl