Hi,

 

I have problem with spinning down of one drive. I have read few topics about, but no solution work for me. I tried to locate problem in logs, but no success.

 

I have Parity (WD30EFRX), Disk1 (ST3000DM001), Disk2 (ST3000DM001) and Disk3 (WD15EARS). Disk3 (sdd) never spin down, others are OK. If I manually spin down all disk in web GUI there is flashing green light near all devices, but if I refresh page, Disk3 has solid green light - so it is spin up.

 

Spinup groups are disabled. I tried both settings to Spin down delay parameter - Use default or concrete time, but the behavior was same, no spin down. I also check command lsof /mnt/* - nothing connected, moreover Disk3 is empty now.

 

One more finding - in SMART disk capabilities is this:

 

Offline data collection status:  (0x82) Offline data collection activity

was completed without error.

Auto Offline Data Collection: Enabled.

 

Could not be this problem? Other disks have this value Disabled. How to disable it?

 

Please point me where I should look, what to try or what could be wrong.

 

Thanks

syslog.txt

I would attach a system log - zip it if it is too large.  You will need that for most anybody to provide real help.

 

However it is usually a plugin or network access that will keep a drive spun up so you could look for things like that.  Maybe disable your plugins and remove network cable from server and see if it will spin down then.

Hi,

 

I have problem with spinning down of one drive. I have read few topics about, but no solution work for me. I tried to locate problem in logs, but no success.

 

I have Parity (WD30EFRX), Disk1 (ST3000DM001), Disk2 (ST3000DM001) and Disk3 (WD15EARS). Disk3 (sdd) never spin down, others are OK. If I manually spin down all disk in web GUI there is flashing green light near all devices, but if I refresh page, Disk3 has solid green light - so it is spin up.

 

Spinup groups are disabled. I tried both settings to Spin down delay parameter - Use default or concrete time, but the behavior was same, no spin down. I also check command lsof /mnt/* - nothing connected, moreover Disk3 is empty now.

 

One more finding - in SMART disk capabilities is this:

 

Offline data collection status:  (0x82) Offline data collection activity

was completed without error.

Auto Offline Data Collection: Enabled.

 

Could not be this problem? Other disks have this value Disabled. How to disable it?

 

Please point me where I should look, what to try or what could be wrong.

 

Thanks

try:

smartctl --offlineauto=off /dev/sdd

 

Are you running any addons or plugins that are installed to that disk?  Some of them will keep a disk spun up.

I attached syslog to first post...

 

I tried your suggestions, but no luck:

 

smartctl --offlineauto=off /dev/sdd - feature is now disabled, but it does not help

 

plugins - I had only SimpleFeatures plugin, but now I have fresh install of Unraid to eliminate problems with plugins - problem is the same

I attached syslog to first post...

 

I tried your suggestions, but no luck:

 

smartctl --offlineauto=off /dev/sdd - feature is now disabled, but it does not help

 

plugins - I had only SimpleFeatures plugin, but now I have fresh install of Unraid to eliminate problems with plugins - problem is the same

I would then install and use the inotify tools package... It will show if any program is accessing the disk.  It is by far the best way.  It is likely one of your PCs on the LAN accessing the disk.

 

See here:

http://lime-technology.com/forum/index.php?topic=25973.msg226486#msg226486

 

Joe L.

I would then install and use the inotify tools package... It will show if any program is accessing the disk.  It is by far the best way.  It is likely one of your PCs on the LAN accessing the disk.

 

See here:

http://lime-technology.com/forum/index.php?topic=25973.msg226486#msg226486

 

Joe L.

 

There are no files on the disk, so nothing should access it. I install inotify tools and this is result, nothing accessing disk shows up:

 

root@Tower:/boot/packages# inotifywait -mr /mnt/disk3

Setting up watches.  Beware: since -r was given, this may take a while!

Watches established.

 

 

 

I would then install and use the inotify tools package... It will show if any program is accessing the disk.  It is by far the best way.  It is likely one of your PCs on the LAN accessing the disk.

 

See here:

http://lime-technology.com/forum/index.php?topic=25973.msg226486#msg226486

 

Joe L.

 

There are no files on the disk, so nothing should access it. I install inotify tools and this is result, nothing accessing disk shows up:

 

root@Tower:/boot/packages# inotifywait -mr /mnt/disk3

Setting up watches.  Beware: since -r was given, this may take a while!

Watches established.

Great, now let it run and see what happens over the next hour or so..  (assuming you've set the spin down timer to 1 hour or less, it should spin down)

 

It monitors in real time.  if you wish to experiment, try accessing the disk from another terminal session or from windows.  You'll then see the accesses.  Since you are probably accessing through user-shares, you might also try monitoring  /mnt/user

I would then install and use the inotify tools package... It will show if any program is accessing the disk.  It is by far the best way.  It is likely one of your PCs on the LAN accessing the disk.

 

See here:

http://lime-technology.com/forum/index.php?topic=25973.msg226486#msg226486

 

Joe L.

 

There are no files on the disk, so nothing should access it. I install inotify tools and this is result, nothing accessing disk shows up:

 

root@Tower:/boot/packages# inotifywait -mr /mnt/disk3

Setting up watches.  Beware: since -r was given, this may take a while!

Watches established.

Great, now let it run and see what happens over the next hour or so..  (assuming you've set the spin down timer to 1 hour or less, it should spin down)

 

It monitors in real time.  if you wish to experiment, try accessing the disk from another terminal session or from windows.  You'll then see the accesses.  Since you are probably accessing through user-shares, you might also try monitoring  /mnt/user

 

I have set up spin down timer to 15 min. I waited for at least 30 mins, no activity in inotify, but no spin down. Other disks spin down correctly

Are you absolutely sure that it is spun up?  Do you have a different way to test, such as metal rod against it to feel for vibration, and cessation of vibration.  Another way is to install and use UnMENU and see what it says.  Or you can try the following hdparm command at a console prompt (parameter case is very important):

hdparm -C /dev/sdX

Replace X with the appropriate letter for that drive.  It should return a response indicating the power mode status of the drive, "standby" indicates spun down.  To manually spin the drive down, you can also use the following command (use -y, do not use -Y or you may need to reboot to regain the drive):

hdparm -y /dev/sdX

Please post your disk.cfg from /flash/config

Are you absolutely sure that it is spun up?  Do you have a different way to test, such as metal rod against it to feel for vibration, and cessation of vibration.  Another way is to install and use UnMENU and see what it says.  Or you can try the following hdparm command at a console prompt (parameter case is very important):

hdparm -C /dev/sdX

Replace X with the appropriate letter for that drive.  It should return a response indicating the power mode status of the drive, "standby" indicates spun down.  To manually spin the drive down, you can also use the following command (use -y, do not use -Y or you may need to reboot to regain the drive):

hdparm -y /dev/sdX

 

Thanks for hdparm command. I tried to play with it and found this:

 

1. if I spin down disks from web interface, there is brief moment when Disk3/sdd is in  standby, but in second it switch to active/idle (checked by hdparm -C /dev/sdd). Other disks stays in standby.

2. if I use command hdparm -y /dev/sdd, disk stays in standby until it is accessed again. But it still will not spin down itself.

Please post your disk.cfg from /flash/config

 

Here it is.

disk.cfg

have you tried using inotifywait on /mnt/user yet?

have you tried using inotifywait on /mnt/user yet?

 

Just did it. I set watches both on /mnt/disk3 and /mnt/user - no activity there and again no spin down of Disk3

have you tried using inotifywait on /mnt/user yet?

 

Just did it. I set watches both on /mnt/disk3 and /mnt/user - no activity there and again no spin down of Disk3

It may be the firmware on that disk just does not like to sleep.

 

Are you sure it still says that the offlineauto=off ?

I've no idea when that parameter is read... might have to power cycle the drive for it to take effect.

 

Joe L.

Please post your disk.cfg from /flash/config

 

Here it is.

 

Thanks, sorry didn't have time to review until now. A few questions:

 

1) What unRAID license are you using? (Basic, Plus or Pro, looks like Plus)

 

2) You are using three drives, #1 - Parity, #2 & #3 are data drives (assign to the array) and no cache drive right?

 

3) Can you explain in detail where are these 3 drives connecting too (ex. All 3 are going to the onboard SATA controller xxx. OR 1 drive (parity) is connected to on board SATA, the other (#2 & #3) are connected to a SIL Sata Controller.)

 

I have an idea and suggestion but want to make sure I understand the answers to the above questions first.

 

have you tried using inotifywait on /mnt/user yet?

 

Just did it. I set watches both on /mnt/disk3 and /mnt/user - no activity there and again no spin down of Disk3

It may be the firmware on that disk just does not like to sleep.

 

Are you sure it still says that the offlineauto=off ?

I've no idea when that parameter is read... might have to power cycle the drive for it to take effect.

 

Joe L.

 

Yes, now it is still offlineauto=off. I tried also power cycle, but same behavior.

Please post your disk.cfg from /flash/config

 

Here it is.

 

Thanks, sorry didn't have time to review until now. A few questions:

 

1) What unRAID license are you using? (Basic, Plus or Pro, looks like Plus)

 

2) You are using three drives, #1 - Parity, #2 & #3 are data drives (assign to the array) and no cache drive right?

 

3) Can you explain in detail where are these 3 drives connecting too (ex. All 3 are going to the onboard SATA controller xxx. OR 1 drive (parity) is connected to on board SATA, the other (#2 & #3) are connected to a SIL Sata Controller.)

 

I have an idea and suggestion but want to make sure I understand the answers to the above questions first.

 

No problem with delay, I appreciate your effort

 

1) Yes, it is Plus license

2) I have four drives: Parity/sda (WD30EFRX), Disk1/sdb (ST3000DM001), Disk2/sdc (ST3000DM001) and Disk3/sdd (WD15EARS), no cache drive

3) I have motherboard ASUS P8H77-I, all disks are connected to it: Parity - SATA6G_1, Disk1 - SATA6G_2, Disk2 - SATA3G_1, Disk3 - SATA3G_2, SATA3G_3 and SATA3G_4 are not connected

OK (thanks for the info), lets start with this:

 

1) Please reboot and do NOT start the array (change to not have the array auto start before you reboot)

 

2) Telnet or via Console; hdparm -C /dev/sdd (if disk 3 is still sdd) this will show you current spin status (should be spun up as you just rebooted)

 

3) Now spin it down; hdparm -y /dev/sdd

 

4) Check status again, as many times as you wish  hdparm -C /dev/sdd (crossing my finger for you that you will be seeing "drive state is:  standby")

 

5) Once you tested it doesnt spin up; hdparm -S0 /dev/sdd (spins it back up)

 

6) Check status again; hdparm -C /dev/sdd (should see "drive state is:  active/idle")

 

Let us know your finding with this test, while the array was stopped.

 

P.S. not sure if you have to undo "smartctl --offlineauto=off" for this test, but go for it as it stands right now.

OK (thanks for the info), lets start with this:

 

1) Please reboot and do NOT start the array (change to not have the array auto start before you reboot)

 

2) Telnet or via Console; hdparm -C /dev/sdd (if disk 3 is still sdd) this will show you current spin status (should be spun up as you just rebooted)

 

3) Now spin it down; hdparm -y /dev/sdd

 

4) Check status again, as many times as you wish  hdparm -C /dev/sdd (crossing my finger for you that you will be seeing "drive state is:  standby")

 

5) Once you tested it doesnt spin up; hdparm -S0 /dev/sdd (spins it back up)

 

6) Check status again; hdparm -C /dev/sdd (should see "drive state is:  active/idle")

 

Let us know your finding with this test, while the array was stopped.

 

P.S. not sure if you have to undo "smartctl --offlineauto=off" for this test, but go for it as it stands right now.

 

OK, here are the results:

 

2) /dev/sdd: drive state is:  active/idle

3) /dev/sdd: issuing standby command

4) /dev/sdd: drive state is:  standby and it does not spin up

5) /dev/sdd: setting standby to 0 (off)

6) /dev/sdd: drive state is:  active/idle

 

It seems OK.

 

But I found one thing - If I do same test with open administrative webpage in browser, then no disk go to standby after hdparm -y. If disks are in standby and I open webpage then they go to active/idle. And if I start array and do the same steps, then disks stay in standby regardles webpage is open.

And last - if I spin down all disks from webpage (array is started) only sda,sdb,sdc stay spun down, sdd is in standby for a moment, but then go active/idle

 

OK, so we know if you just load unRAID with nothing going on (array stopped, etc...) the drive itself is more than capable of spinning down and maintaining that state.

So that good news for you and nothing magical about the drive.

 

But I found one thing - If I do same test with open administrative webpage in browser, then no disk go to standby after hdparm -y.

You lost me here, what is a "administrative webpage in browser?

 

If disks are in standby and I open webpage then they go to active/idle. And if I start array and do the same steps, then disks stay in standby regardles webpage is open.

And last - if I spin down all disks from webpage (array is started) only sda,sdb,sdc stay spun down, sdd is in standby for a moment, but then go active/idle

Sorry, again open what web page this time (seems by your wording your opening 2 different web pages)?

 

And last - if I spin down all disks from webpage (array is started) only sda,sdb,sdc stay spun down, sdd is in standby for a moment, but then go active/idle

Right, the continued (original) behaviour and why you posted here. I didn't say my steps to test were going to solve your problem, we needed to take 2 steps back and make sure it was not a hard drive issue first (one that refuses to spin-down and stay spun down).

 

Going back to your disk.cfg file.

Based on the configuration you stated, I am taking a stab that each onboard port is issued an individual host# so

ParityDrive = "host0"

Disk#1 = "host1"

Disk#2 = "host2"

Disk#3 = "host3"

 

Where with controller cards, say an 8 port card having 8 disks attached to it, each of the 8 disks is assign to the single host# the controller was issued, say Host#2 (hope that makes sense. Disk#1/2/3/4/5/6/7/8, all = "Host#2" as an example)

 

So the only thing I can see is disk#3's value diskSpindownDelay.3="15" should be "-1" as you have a global spindowndelay of "15" and spinupgroups equaling "no", so again I am going to guess you manually clicked on Disk#3 and set "Spin down delay" to "15 minutes", so just set it back to "Use Default" (this should set the value to "-1" in disk.cfg). click apply and reboot the box. Test spinning down and see if it stay's down.

 

I am leaning to, you have to have something left over from a plugin at this point that is causing this. Can you make a copy of your USB flash less the license key and drop a copy of it on dropbox or other cloud service for a look. (you can PM me the link)

 

ALSO

 

I see there is a BIO's update available for your board "P8H77-I BIOS Version 0904 ---- Improve system stability." http://www.asus.com/Motherboard/P8H77I/#support_Download_30

 

Hope you have  AHCI enabled (not RAID) in the bios. Also check if there is an option to select AHCI 2.0 (versus 3.0).

Also in the bios if you have something like "fan controlled from operating system", disable and change it to a fixed value (like 50% or 100%, whatever you like but a fixed value).

 

I bring these points up because of these two lines below: (something VERY similar to yours was reported https://bugzilla.kernel.org/show_bug.cgi?id=43059, written off as bad BIOS by manufacturer, thus I would try to update the bios to the last version they have.)

Feb 26 11:15:58 Tower kernel: ACPI Error: [DSSP] Namespace lookup failure, AE_NOT_FOUND (20120320/psargs-359)
Feb 26 11:15:58 Tower kernel: ACPI Error: Method parse/execution failed [\_SB_.PCI0.SAT0.SPT2._GTF] (Node f5060318), AE_NOT_FOUND (20120320/psparse-536)

 

They also mentioned to turn ACPI off altogether, but I would leave this for a later time. Post you syslog of a clean boot, once u have updated your bios and any of these changes you have made in the Bios. Of coarse test the spin down after capturing the syslog for us.

 

But I found one thing - If I do same test with open administrative webpage in browser, then no disk go to standby after hdparm -y.

You lost me here, what is a "administrative webpage in browser?

 

I mean http://tower/Main

 

If disks are in standby and I open webpage then they go to active/idle. And if I start array and do the same steps, then disks stay in standby regardles webpage is open.

And last - if I spin down all disks from webpage (array is started) only sda,sdb,sdc stay spun down, sdd is in standby for a moment, but then go active/idle

Sorry, again open what web page this time (seems by your wording your opening 2 different web pages)?

 

Same page as above

Going back to your disk.cfg file.

Based on the configuration you stated, I am taking a stab that each onboard port is issued an individual host# so

ParityDrive = "host0"

Disk#1 = "host1"

Disk#2 = "host2"

Disk#3 = "host3"

 

Where with controller cards, say an 8 port card having 8 disks attached to it, each of the 8 disks is assign to the single host# the controller was issued, say Host#2 (hope that makes sense. Disk#1/2/3/4/5/6/7/8, all = "Host#2" as an example)

 

So the only thing I can see is disk#3's value diskSpindownDelay.3="15" should be "-1" as you have a global spindowndelay of "15" and spinupgroups equaling "no", so again I am going to guess you manually clicked on Disk#3 and set "Spin down delay" to "15 minutes", so just set it back to "Use Default" (this should set the value to "-1" in disk.cfg). click apply and reboot the box. Test spinning down and see if it stay's down.

 

I am leaning to, you have to have something left over from a plugin at this point that is causing this. Can you make a copy of your USB flash less the license key and drop a copy of it on dropbox or other cloud service for a look. (you can PM me the link)

 

ALSO

 

I see there is a BIO's update available for your board "P8H77-I BIOS Version 0904 ---- Improve system stability." http://www.asus.com/Motherboard/P8H77I/#support_Download_30

 

Hope you have  AHCI enabled (not RAID) in the bios. Also check if there is an option to select AHCI 2.0 (versus 3.0).

Also in the bios if you have something like "fan controlled from operating system", disable and change it to a fixed value (like 50% or 100%, whatever you like but a fixed value).

 

I bring these points up because of these two lines below: (something VERY similar to yours was reported https://bugzilla.kernel.org/show_bug.cgi?id=43059, written off as bad BIOS by manufacturer, thus I would try to update the bios to the last version they have.)

Feb 26 11:15:58 Tower kernel: ACPI Error: [DSSP] Namespace lookup failure, AE_NOT_FOUND (20120320/psargs-359)
Feb 26 11:15:58 Tower kernel: ACPI Error: Method parse/execution failed [\_SB_.PCI0.SAT0.SPT2._GTF] (Node f5060318), AE_NOT_FOUND (20120320/psparse-536)

 

They also mentioned to turn ACPI off altogether, but I would leave this for a later time. Post you syslog of a clean boot, once u have updated your bios and any of these changes you have made in the Bios. Of coarse test the spin down after capturing the syslog for us.

 

Thank you for suggestions. I'll definitely try BIOS settings and post logs and copy of flash on sunday - I'm out of my NAS now.