Re: unRAID Server Release 6.0-beta12-x86_64 SPINDOWN/SPINUP ISSUES


209 posts in this topic Last Reply

Recommended Posts

One other thing that I might try is to put disk 9 on a motherboard SATA port instead of the Supermicro HBA to see if that makes a difference, but I don't know if that would help point you guys in any useful direction.

 

That didn't help either. I didn't think it would but it was worth trying.

 

Gary

Link to post
  • Replies 208
  • Created
  • Last Reply

Top Posters In This Topic

Top Posters In This Topic

Posted Images

If I disable WD Idle I cannot reproduce the problem, so it seems like the issue is caused by an interaction between WD Idle and the unRAID auto spin down sequence.

 

Is this the problem with drives not spinning down?

I'm not sure what you're asking here. If you're asking me to clarify, yes I was writing about the problem with drives not spinning down.

 

If you are asking a broader question then you're probably looking for an answer from others, but I'll give my thoughts anyway. 

 

I think there are probably multiple different spin up/down issues that have been documented in this thread and at least some of them are due to clients or applications accessing files on the array. I also think there is a problem with drive spin down that is related to WD Idle and the unRAID auto spin down sequence and this problem started with v6beta12. Disabling WD Idle is a viable solution for me, but I don't think Lime Technology would want to specify it as a solution for all users with this issue.

Link to post

If I disable WD Idle I cannot reproduce the problem, so it seems like the issue is caused by an interaction between WD Idle and the unRAID auto spin down sequence.

 

Is this the problem with drives not spinning down?

 

wdidle3 is a utility to disable parking on your hard drives. A few people have said that after using this utility to disable parking, their drives now spin down. However, this is not a solution as there are others that have this problem with hard drivers from other manufacturers. Plus, this has only been a problem since beta12 I think.

 

Gary

Link to post

Poking around my server tonight and happened to notice... drives are all spun down. Other than disk1 (I woke it up looking at folder structure) & cache (Plex running), ALL the rest are napping. Will check back again in an hour when all but cache should be spun down.

 

Beta12 I had 3 or 4 drives that would be up and spinning for no reason I could see.

Link to post

Poking around my server tonight and happened to notice... drives are all spun down. Other than disk1 (I woke it up looking at folder structure) & cache (Plex running), ALL the rest are napping. Will check back again in an hour when all but cache should be spun down.

 

Beta12 I had 3 or 4 drives that would be up and spinning for no reason I could see.

I think that with beta 14, a number of folks had this issue resolved, but not all.

 

For those that are still experiencing it, I need to see steps to recreate it again. I'd also be interested if the folks that remain with the issue are only seeing it on WD drives.

Link to post

I think that with beta 14, a number of folks had this issue resolved, but not all.

 

For those that are still experiencing it, I need to see steps to recreate it again. I'd also be interested if the folks that remain with the issue are only seeing it on WD drives.

 

My array:

Parity ST3000DM001-1CH166_W1F1WJCM (sdb)

Disk 1 ST3000DM001-1CH166_W1F41LSA (sdd)

Disk 2 ST3000DM001-1CH166_Z1F4CAS0 (sde)

Disk 3 ST3000DM001-1ER166_W500JBH3 (sdf)

Disk 4 WDC_WD20EARS-00MVWB0_WD-WMAZA3888280 (sdg)

Disk 5 ST3000DM001-1CH166_Z1F4AP89 (sdl)

Disk 6 ST3000DM001-1ER166_Z5007S1K (sdk)

Disk 7 ST3000DM001-1CH166_W1F1WJ9Y (sdj)

Disk 8 ST3000DM001-1CH166_W1F1WHCQ (sdi)

Disk 9 WDC_WD20EARS-00MVWB0_WD-WCAZA4999014 (sdm)

Cache WDC_WD6400AAKS-00A7B0_WD-WMASY1070643 (sdc)

 

Mounted outside the array for VM's:

SAMSUNG_SSD_830_Series_S0WFNYABC00539 (sdh)

 

No plugins installed, no unMenu, clean go file, running under Xen with 1 Windows VM running.

 

Disk 9 is the drive that will not spin down. It is the only effected disk since beta 12. Both disk 4 and disk 9 have the same firmware v51.0AB51. Disk 4 is on the motherboard SATA port and disk 9 is on a Supermicro AOC-SASLP-MV8 PCI-E x4 HBA. I tried swapping the SATA ports between disk 4 and disk 9 but it didn't make any difference as disk 9 still doesn't spin down like it should.

 

Each disk in the array is on the same spinup group using the default spin down timer set at 2 hours. Backups from Windows PC's run nightly starting at 2:30 AM and end a few hours later and then the drives spin down around 6:00 AM. Other than that, I access the server 2 or 3 days a week to watch movies. The server should be inactive most of the day.

 

When the server is inactive for 2 hours or so, all drives spin down including cache and parity, except for disk 9. I can click the Spin Down button under Main -> Array Operation and disk 9 will then spin down and it stays spun down until the array is accessed again. Another oddity is that when the array is spun up, I can change the Default Spin Down Delay under Settings -> Disk Settings, and the array (including disk 9) will then spin down after the new delay setting expires. It will stay spun down until accessed again. This behavior only works one time though. After the array is accessed again and spun up, disk 9 no longer spins down again like the rest of the array does. I can repeat this behavior of changing the Default Spin Down Delay after the array is spun up, and disk 9 will spin down like the rest of the array does.

 

Gary

syslog.txt

Link to post

I notice - including myself - that it is usually the last disk in the array which has an issue.

 

My observation: all disks spin down after the spun-down timer expires, but almost always the last disk spins-up immediately, and stays up. Sometimes the behaviour also happens to some of my other disks in the array (it is a bit random).

 

In my case disk 10 (the last disk) is a WD with a different firmware version as compared to the other WDs in the array. I rebuild the array and swapped disk 9 and 10 to see whether the issue moves with the disk, it doesn't.

 

This was tested with page refresh enabled and disabled, and with notifications on and off, it does not make a difference as far as I can see.

 

 

Link to post

I notice - including myself - that it is usually the last disk in the array which has an issue.

 

My observation: all disks spin down after the spun-down timer expires, but almost always the last disk spins-up immediately, and stays up. Sometimes the behaviour also happens to some of my other disks in the array (it is a bit random).

 

In my case disk 10 (the last disk) is a WD with a different firmware version as compared to the other WDs in the array. I rebuild the array and swapped disk 9 and 10 to see whether the issue moves with the disk, it doesn't.

 

This was tested with page refresh enabled and disabled, and with notifications on and off, it does not make a difference as far as I can see.

 

Interesting. I was wondering if this problem could be tied in with being the last disk in the array. I can also confirm that page refresh setting disabled and notifications disabled doesn't make any difference.

 

Gary

Link to post

 

For those that are still experiencing it, I need to see steps to recreate it again. I'd also be interested if the folks that remain with the issue are only seeing it on WD drives.

 

It happens in my system with a WD40EZRX drive in the parity slot.

 

Steps to recreate:

[*]Spin up the problem drive by itself

[*]Wait for spin down delay

[*]"spindown 0" appears in the syslog, but drive does not spin down

 

I watched power consumption and listened as "spindown 0" appeared in the syslog.  I heard disk head movement but power consumption did not drop as it usually does with a successful spin down.

Link to post

 

For those that are still experiencing it, I need to see steps to recreate it again. I'd also be interested if the folks that remain with the issue are only seeing it on WD drives.

 

It happens in my system with a WD40EZRX drive in the parity slot.

 

Steps to recreate:

[*]Spin up the problem drive by itself

[*]Wait for spin down delay

[*]"spindown 0" appears in the syslog, but drive does not spin down

 

I watched power consumption and listened as "spindown 0" appeared in the syslog.  I heard disk head movement but power consumption did not drop as it usually does with a successful spin down.

 

Unfourtunately, for limetech, it hasn't been this easy to reproduce. They were able to in beta12, then moved to 13 and saw the issue go away. But based on what jonp said earlier, possible we had two different issues with one solved now (non-WD drives).

 

I had all WD drives for beta12, but recently added a Seagate drive I wasn't using for anything else.

 

The Seagate drive doesn't have content on it yet, but I plan to use it as an archive disk for TV Shows I don't re-watch often.

 

Currently for me, data disk 1 (WDC_WD20EARX-00PASB0) and 2 (WDC_WD20EARS-00MVWB0) are the ones that aren't spinning down. Both WD green drives (all mine are green).

 

I didn't have any luck in beta12 tracking this down, so not sure what to do next. I also have notifications disabled and UI refresh disabled.

Link to post

Unfourtunately, for limetech, it hasn't been this easy to reproduce. They were able to in beta12, then moved to 13 and saw the issue go away. But based on what jonp said earlier, possible we had two different issues with one solved now (non-WD drives).

 

I didn't have any luck in beta12 tracking this down, so not sure what to do next. I also have notifications disabled and UI refresh disabled.

 

Yeah, I know. Personally, I should just replace this drive or not worry about it, but that doesn't really help anyone. But, if this issue persists through RC or Release, then the people that might experience this problem could increase quite a bit. So, I'll do what I can to help.

 

On the other hand, it might be best to just push this issue aside rather than waste a lot of resources on it. Who knows, it might just resolve itself over the next few versions anyways.

 

Gary

Link to post

Unfourtunately, for limetech, it hasn't been this easy to reproduce. They were able to in beta12, then moved to 13 and saw the issue go away. But based on what jonp said earlier, possible we had two different issues with one solved now (non-WD drives).

 

I didn't have any luck in beta12 tracking this down, so not sure what to do next. I also have notifications disabled and UI refresh disabled.

 

Yeah, I know. Personally, I should just replace this drive or not worry about it, but that doesn't really help anyone. But, if this issue persists through RC or Release, then the people that might experience this problem could increase quite a bit. So, I'll do what I can to help.

 

On the other hand, it might be best to just push this issue aside rather than waste a lot of resources on it. Who knows, it might just resolve itself over the next few versions anyways.

 

Gary

 

Well you guys are in luck, I'm not sure how I did it but I just recreated this again and it is affecting both a WD and a Seagate disk.  This was on beta 14.  I just upgraded this test system to 14a, adjusted my spin down timers to 15 minutes, and I'm anxiously awaiting the results.

Link to post

Your 15 minutes are up.

 

I think we're just looking at this wrong. This isn't a bug, it's a feature.

 

"unRAID Predictive Read and Write Technology© that automagically uses advanced adaptive algorithms to prevent hard drives from spinning down that you need to read and write data to in the not so distant future."

 

Gary

Link to post

Your 15 minutes are up.

 

I think we're just looking at this wrong. This isn't a bug, it's a feature.

 

"unRAID Predictive Read and Write Technology© that automagically uses advanced adaptive algorithms to prevent hard drives from spinning down that you need to read and write data to in the not so distant future."

 

Gary

 

WOOT!!!

 

spindown-success.png

Link to post

Ok, I can confirm that a once-affected system by this bug that had two disks (a WD and a Seagate) which weren't spinning down are spinning down per policy settings in beta 14a.  I'll be very anxious to hear feedback from those in this thread once we release 14a and folks upgrade.

 

set fingers_crossed 1

Link to post

This indeed sounds promising, but I am silently wondering how you can be sure it was 14a that solved the issue on your test system if you were actually unsure previously how you reproduced the issue with 14? :)

 

I had the same thought. In addition, not know what fixed it if it actually is this time, could lead to another break in the future.

Link to post

This indeed sounds promising, but I am silently wondering how you can be sure it was 14a that solved the issue on your test system if you were actually unsure previously how you reproduced the issue with 14? :)

Not so silently wondering. :)

What changed from 14 to 14a that fixed the issue? If you don't know, how will you be able to keep the problem from coming back in future releases?

Link to post

This indeed sounds promising, but I am silently wondering how you can be sure it was 14a that solved the issue on your test system if you were actually unsure previously how you reproduced the issue with 14? :)

 

I'm going to revert to beta14 on that test system in a little bit here to see if I can recreate it again there, then try upgrading again to 14a to make sure.

 

As far as what fixed this, it could very well be the change we made in 14a to poll_spindown to not tie that event to checking disk attributes (SMART).

Link to post

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.

Guest
Reply to this topic...

×   Pasted as rich text.   Restore formatting

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.