[Plugin] Spin Down SAS Drives

[Dustiebin]

On 5/25/2022 at 5:57 AM, another_hoarder said:

Sweet, thanks @doron - looking forward to some debugging fun

Hi @another_hoarder

did you work out what was causing your spinups? 

 

I am having a similar issue.

I spin them down and within minutes they are back up and running again. 

I have closed down every docker running and removed 'apps' that I read could spin up the drives but still getting it happen. 

 

It has only happened since my new build and I moved my drives to it. Big difference is that they are now plugged into an HBA card (HP 240). They are SATA drives with one SAS drive. The SAS drive is the only one that stays spun down. 

 

thanks

[SimonF]

2 hours ago, Dustiebin said:

Hi @another_hoarder

did you work out what was causing your spinups? 

 

I am having a similar issue.

I spin them down and within minutes they are back up and running again. 

I have closed down every docker running and removed 'apps' that I read could spin up the drives but still getting it happen. 

 

It has only happened since my new build and I moved my drives to it. Big difference is that they are now plugged into an HBA card (HP 240). They are SATA drives with one SAS drive. The SAS drive is the only one that stays spun down. 

 

thanks

This plugin only adds support for SAS to the existing SATA support included in stock.

 

hdparm -C /dev/sdx is used to check status and hdparm -y for spin down on SATA drives.

 

What output do you get for those.

 

also what output do you get for smartctl -n standby /dev/sdx?

 

Maybe worth starting a general thread and post Diagnostics.

[another_hoarder]

On 6/4/2022 at 3:30 PM, Dustiebin said:

did you work out what was causing your spinups? 

Hi @Dustiebin

 

Unfortunately, I did not.  Doron's utility points out that the parent process is emhttpd but I can't see anything upwards of that.  I tried stopping all dodckers as well to no avail, and have tried turning off plugins, also with no luck.  I think I've combed through most solutions proposed to this process (dynamix plugins, nerdpack tweaks, killing dockers, booting fresh), and nada.

 

I'm very hopeful that 6.10.x will fix it since there's some changes to how smartctl works in that version, but so far I'm resisting the urge to upgrade given how many people seem to have had issues.  Just don't have the time to play with it ftm.

[Dustiebin]

On 6/4/2022 at 11:00 PM, SimonF said:

This plugin only adds support for SAS to the existing SATA support included in stock.

 

hdparm -C /dev/sdx is used to check status and hdparm -y for spin down on SATA drives.

 

What output do you get for those.

 

also what output do you get for smartctl -n standby /dev/sdx?

 

Maybe worth starting a general thread and post Diagnostics.

Hi @SimonF Thanks for taking the time to read my post and your ideas.

 

I did the hdparm -C and all SATA drives stated up. 

I did the hdparm -y and all drives show as spun down --- funny thing is Unraid still sees them as ACTIVE. Logs of the drives do not show the spin down I did via hdparm.... 

 

when I then do the smartctl option it tells me they are ACTIVE or IDLE.....

 

Sounds like I have something messed up moving it from my old micro HP server to my new rig and my HP240 card.... 

[Dustiebin]

On 6/6/2022 at 6:27 PM, another_hoarder said:

Hi @Dustiebin

 

Unfortunately, I did not.  Doron's utility points out that the parent process is emhttpd but I can't see anything upwards of that.  I tried stopping all dodckers as well to no avail, and have tried turning off plugins, also with no luck.  I think I've combed through most solutions proposed to this process (dynamix plugins, nerdpack tweaks, killing dockers, booting fresh), and nada.

 

I'm very hopeful that 6.10.x will fix it since there's some changes to how smartctl works in that version, but so far I'm resisting the urge to upgrade given how many people seem to have had issues.  Just don't have the time to play with it ftm.

Thanks for gettting back to me @another_hoarder

I have upgraded to 6.10.2 and still the same issues.......

[SimonF]

25 minutes ago, Dustiebin said:

funny thing is Unraid still sees them as ACTIVE. Logs of the drives do not show the spin down I did via hdparm...

Yes system wont know they are spundown if done outside of gui.

[SimonF]

47 minutes ago, Dustiebin said:

Hi @SimonF Thanks for taking the time to read my post and your ideas.

 

I did the hdparm -C and all SATA drives stated up. 

I did the hdparm -y and all drives show as spun down --- funny thing is Unraid still sees them as ACTIVE. Logs of the drives do not show the spin down I did via hdparm.... 

 

when I then do the smartctl option it tells me they are ACTIVE or IDLE.....

 

Sounds like I have something messed up moving it from my old micro HP server to my new rig and my HP240 card.... 

Is your hp240 in  hba mode?

[Dustiebin]

20 hours ago, SimonF said:

Is your hp240 in  hba mode?

I downloaded the HP240 boot program to one get it up to date on firmware and then to put it in HBA mode. 

This was before I put it in the unraid box.

I am not sure now how to tell if it is in HBA mode to double check. 

Is it possible in Unraid or do I need to pull it and put it in another box? 

 

thanks for your help by the way. 

[SimonF]

8 minutes ago, Dustiebin said:

I downloaded the HP240 boot program to one get it up to date on firmware and then to put it in HBA mode. 

This was before I put it in the unraid box.

I am not sure now how to tell if it is in HBA mode to double check. 

Is it possible in Unraid or do I need to pull it and put it in another box? 

 

thanks for your help by the way. 

I did find a video where they use ssacli to look at the card. It wont be in stock unraid, will see if I can find a version that can be downloaded and will work on unraid. But may take me a few days.

[DavidNguyen]

Hello, thank you for a very useful plugin.

I've just installed 3 HGST SAS drives into what will be an 8 drive array, currently attached via a Dell HBA card.

Currently they will spindown when hitting the button in the GUI, only for them to spin back up again a few minutes later. This is despite the GUI showing that there is no data being written or read.

 

All models are HUS724040ALS640 4TB drives.

 

Cheers if this can be sorted out for a quiter system.

[doron]

Currently they will spindown when hitting the button in the GUI, only for them to spin back up again a few minutes later.

Spinning up a few minutes later would probably be due to some activity against the drive.
You may want to check for other plugins, Dockers or VMs that generate periodical I/O against the array.

Sent from my tracking device using Tapatalk

[SimonF]

30 minutes ago, DavidNguyen said:

Hello, thank you for a very useful plugin.

I've just installed 3 HGST SAS drives into what will be an 8 drive array, currently attached via a Dell HBA card.

Currently they will spindown when hitting the button in the GUI, only for them to spin back up again a few minutes later. This is despite the GUI showing that there is no data being written or read.

 

All models are HUS724040ALS640 4TB drives.

 

Cheers if this can be sorted out for a quiter system.

In addition to the points doron has said which mode do you have set? Maybe worth posting diagnostics.

 

[DavidNguyen]

1 hour ago, doron said:

Spinning up a few minutes later would probably be due to some activity against the drive.
You may want to check for other plugins, Dockers or VMs that generate periodical I/O against the array.

 

I don't think it was this, considering that the drives have just been cleared, formatted and are sitting empty.

But as a test, I disabled all my containers & VMs, spun down all the drives and left it to idle.

After a few minutes, my SAS drives spun up again, even before the Parity or Appdata drives had done so.

Having said that, all disks in the array have now become active as I write this.

 

I've attached my diagnostics file below as requested.

 

EDIT: after playing around and deleting more plugins, I think it was the Dyanamix Autofan that was waking up the drives. Just did a search and looks like it has been encountered before, including a possible workaround.  I'll keep playing around to see what works best for me. Cheers for your input though guys

tower-diagnostics-20220618-1835.zip

Edited June 18, 2022 by DavidNguyen

[doron]

51 minutes ago, DavidNguyen said:

I've attached my diagnostics file below as requested.

Just for the sake of testing, can you disable the System Autofan and System Temp plugins and retry?

 

EDIT: Reading the code of System Autofan, it will definitely spin up SAS drives every $INTERVAL minutes. My guess is that if you disable this plugin, the phenomenon you describe will go away.

I will contact the author.

Edited June 18, 2022 by doron

[DavidNguyen]

5 hours ago, doron said:

I will contact the author.

That's great, thank you. I'll follow this thread; please keep us updated and see if collaboration can be made, as both plugins have very useful functions.

[doron]

@DavidNguyen , meanwhile please keep us posted as to whether disabling the autofan plugin indeed stopped the phenomenon.

If you want to test, I can send you a (hacked, unauthorized) code fix for that plugin to try on.

Sent from my tracking device using Tapatalk

[DavidNguyen]

9 hours ago, doron said:

please keep us posted as to whether disabling the autofan plugin indeed stopped the phenomenon.
 

Yeah I left it overnight with the plugin enabled, just ignoring the SAS drives in question. I've woken up to them still in standby mode.

I'd be open to testing a modified version of either plugin(s)

[doron]

Okay @DavidNguyen, here goes.

 

Again, please note this is not sanctioned by the plugin author and should be considered a hack, just for testing.

Also note that I didn't package it back into plugin format, which means that the change will be lost upon reboot.

 

Attached please find the modified script "autofan".

 

Steps to activate:

- Place the script in /usr/local/emhttp/plugins/dynamix.system.autofan/scripts/

- chmod 755 autofan

- Go to settings, Fan Auto Control, set the function to "Disable", then Apply

- (Make sure process "autofan" is not running)

- Set function back to "Enable", then Apply

 

You should now have autofan function without offending your SAS drives.

 

Please report success / issues.

autofan

[DavidNguyen]

On 6/19/2022 at 5:28 PM, doron said:

Please report success / issues.

 

Looking good so far. I've been transferring around 10TB over the 3 SAS drives, they are interacting with the fan-speed plugin, and they are not being woken up when spun down.

[doron]

10 minutes ago, DavidNguyen said:

 

Looking good so far. I've been transferring around 10TB over the 3 SAS drives, they are interacting with the fan-speed plugin, and they are not being woken up when spun down.

Very good. Now let's hope @bonienl considers adding the code fix into the plugin.

[DavidNguyen]

On 6/19/2022 at 5:28 PM, doron said:

Steps to activate:

- Place the script in /usr/local/emhttp/plugins/dynamix.system.autofan/scripts/

- chmod 755 autofan

- Go to settings, Fan Auto Control, set the function to "Disable", then Apply

- (Make sure process "autofan" is not running)

- Set function back to "Enable", then Apply

On 6/22/2022 at 5:47 PM, doron said:

Very good. Now let's hope @bonienl considers adding the code fix into the plugin.

How/where might I create a bit of a script to do this at boot in the meantime?

 

[SimonF]

4 hours ago, DavidNguyen said:

How/where might I create a bit of a script to do this at boot in the meantime?

 

You can add to the /boot/config/go file before emhttp &

[drc13]

Hi

Hopefully this is the correct thread to ask for advice.

I currently have a 11 disk array with a combination of SATA and SAS drives. Spin down works perfectly except for any SAS drives with model number ST2000NM0001 they won't spin down when triggered either automatically or manually. All my other SATA and SAS drives spin down and stay spun down perfectly using this plugin.

I replaced 2 of the ST2000NM0001 with alternate drives and they immediately worked with spinning down so I now have 1 remaining ST2000NM0001 that is the only drive in my array spinning permanently

With some guidance I'm open to doing some troubleshooting and I've attached the diagnostics log.

Thanks in advance

tower-diagnostics-20220707-0727.zip

[jeffrey.el]

Hi

I'm hoping maybe someone can help me out or maybe explain why I can't seem to get my SAS drives to spin down..

I currently have 3 SAS drives in my Unraid server, all the same model; WDC WD3001FYYG-0

Whenever a disk is idle or when I manually spin down one of these disks, I see a message in the log "SAS Assist v2022.05.25: Spinning down device"

 

So it seems the plugin is working! However, the activity led on the disks remain lit which would indicate that the disks are still spun up.. 

 

Is there anyone who might have an idea what is causing this?

[SimonF]

10 hours ago, jeffrey.el said:

Hi

I'm hoping maybe someone can help me out or maybe explain why I can't seem to get my SAS drives to spin down..

I currently have 3 SAS drives in my Unraid server, all the same model; WDC WD3001FYYG-0

Whenever a disk is idle or when I manually spin down one of these disks, I see a message in the log "SAS Assist v2022.05.25: Spinning down device"

 

So it seems the plugin is working! However, the activity led on the disks remain lit which would indicate that the disks are still spun up.. 

 

Is there anyone who might have an idea what is causing this?

can you run sdparm -C sense /dev/sdx with the correct drive for the 3 WDCs once you think the drives are spun down?

 

or you can do smartctl -n standby /dev/sdx

 

 

What output to you get?

Edited July 9, 2022 by SimonF